mirror of
https://github.com/kennethreitz/tablib.git
synced 2026-06-05 06:56:13 +00:00
Adding DBF support.
Squashing two squashes. Adding DBF support Adding the DBFpy python package The DBFpy package provides basic dbf support for python. Still need to write an interface format file for tablib. Adding DBF format and imports in compat.py Adding DBF format to formats.__init__ DBF format had not been committed to formats.__init__, so I’m adding it. Adding a dbf import test Adding at test to check whether a DBF can be created properly and compare it against a regression binary string. Adding an import_set test (and renaming another) Adding an import_set test that conforms with the other import_set tests for other formats. I’m also adding an export_set function. Fixing system site-packages import Importing dbfpy from tab lib.packages instead of system site packages. Fixing a syntaxError in dbfpy/dbfnew.py Fixing an issue with ending field definitions DBFPY, when writing a DBF, terminates the field definitions with a newline character. When importing a DBF from a stream, however, DBFPY was looking only for the \x0D character rather than the newline. Now we consider both cases. Adding a test for dbf format detection Adding DBF filetype detection tests Adding tests for YAML, JSON, TSV, CSV using the DBF detection function. Handling extra exceptions in dbf detection Adding exception handling for struct.error, an exception that DBFPY raises when trying to unpack a TSV table. Since it’s not a DBF file, we know it’s not a DBF and return False. Fixing an issue with the DBF set exporting test The DBF set export test needed a bit enabled (probably the writeable bit?) before the test would match the regression output. Updating dbf interface Updating the int/float class/type checking in the dbf format file. This allows for python2 and python3 compatibility. Tweaking dbfpy to work with python3 Altering a couple of imports. Updating dbf tests for binary data compatibility Making regression strings binary and improving debug messages for dbf assertion errors. Improving file handling for python 2 and 3 Updating DBF file handling for both python 2 and 3 in the _dbf interface. Adding a (seemingly) functional dbfpy for python3 I’ve made dbfpy python3 compatible! Tests appear to pass. A significant change was made to the format detection test whereby I made the input string a binary (bytes) string. If the string is not a bytes string by the time we try to detect the format, we try to decode the string as utf-8 (which admittedly might not be the safest thing to do) and try to decode anyways. Updating imports for tablib dbf interface Now importing python2 or python3 versions as appropriate. Updating dbf package references in compat.py Cleaning up debugging print statements Updating stream handling in dbf interface Factoring the open() call out of the py3 conditional and removing the temp file before returning the stream value. Adding dbfpy3 init.py I had apparently missed the dbfpy3 init file when committing dbfpy3. Adding dbfpy and dbfpy3 to setup.py's package list Switching test order of formats Putting dbf format testing ahead of TSV. In some of my tests with numeric DBF files, I encountered an issue where the ASCII horizontal tab character (0x09) would appear in a numeric DBF. Because of the order of tabular format imports, though, format detection would recognize it as a TSV and not as a DBF. Adding my name to AUTHORS. Adding a DBF property to tab lib core Documentation includes examples on how to explicitly load a DBF straight from a file and how to load a DBF from a binary string. Also, how to write the binary data to a file. Adding DBF format notes to README Adding exclamation point to DBF section title Matching formatting of XLS section Updating setup.py to match current dev state Setup.py had been updated since I forked the tablib repo, so I’m updating setup.py to match its current structure while still maintaining DBF compatibility. Fixed callable collumn test the test was sending a list instead of a function CORE CONTRIBUTORS 🍰 @iurisilvio v0.10.0 WHEELS 3.3, 3.4 makefile for WHEELS v0.10.0 history ALL Separate py2 and py3 packages to avoid installation errors. Fix #151 Running travis and tox with python 3.4. Adding DBF support Adding the DBFpy python package The DBFpy package provides basic dbf support for python. Still need to write an interface format file for tablib. Adding DBF format and imports in compat.py Adding DBF format to formats.__init__ DBF format had not been committed to formats.__init__, so I’m adding it. Adding a dbf import test Adding at test to check whether a DBF can be created properly and compare it against a regression binary string. Adding an import_set test (and renaming another) Adding an import_set test that conforms with the other import_set tests for other formats. I’m also adding an export_set function. Fixing system site-packages import Importing dbfpy from tab lib.packages instead of system site packages. Fixing a syntaxError in dbfpy/dbfnew.py Fixing an issue with ending field definitions DBFPY, when writing a DBF, terminates the field definitions with a newline character. When importing a DBF from a stream, however, DBFPY was looking only for the \x0D character rather than the newline. Now we consider both cases. Adding a test for dbf format detection Adding DBF filetype detection tests Adding tests for YAML, JSON, TSV, CSV using the DBF detection function. Handling extra exceptions in dbf detection Adding exception handling for struct.error, an exception that DBFPY raises when trying to unpack a TSV table. Since it’s not a DBF file, we know it’s not a DBF and return False. Fixing an issue with the DBF set exporting test The DBF set export test needed a bit enabled (probably the writeable bit?) before the test would match the regression output. Updating dbf interface Updating the int/float class/type checking in the dbf format file. This allows for python2 and python3 compatibility. Tweaking dbfpy to work with python3 Altering a couple of imports. Updating dbf tests for binary data compatibility Making regression strings binary and improving debug messages for dbf assertion errors. Improving file handling for python 2 and 3 Updating DBF file handling for both python 2 and 3 in the _dbf interface. Adding a (seemingly) functional dbfpy for python3 I’ve made dbfpy python3 compatible! Tests appear to pass. A significant change was made to the format detection test whereby I made the input string a binary (bytes) string. If the string is not a bytes string by the time we try to detect the format, we try to decode the string as utf-8 (which admittedly might not be the safest thing to do) and try to decode anyways. Updating imports for tablib dbf interface Now importing python2 or python3 versions as appropriate. Updating dbf package references in compat.py Cleaning up debugging print statements Updating stream handling in dbf interface Factoring the open() call out of the py3 conditional and removing the temp file before returning the stream value. Adding dbfpy3 init.py I had apparently missed the dbfpy3 init file when committing dbfpy3. Adding dbfpy and dbfpy3 to setup.py's package list Switching test order of formats Putting dbf format testing ahead of TSV. In some of my tests with numeric DBF files, I encountered an issue where the ASCII horizontal tab character (0x09) would appear in a numeric DBF. Because of the order of tabular format imports, though, format detection would recognize it as a TSV and not as a DBF. Adding my name to AUTHORS. Adding a DBF property to tab lib core Documentation includes examples on how to explicitly load a DBF straight from a file and how to load a DBF from a binary string. Also, how to write the binary data to a file. Adding DBF format notes to README Adding exclamation point to DBF section title Matching formatting of XLS section Updating setup.py to match current dev state Setup.py had been updated since I forked the tablib repo, so I’m updating setup.py to match its current structure while still maintaining DBF compatibility. Fixed callable collumn test the test was sending a list instead of a function CORE CONTRIBUTORS 🍰 @iurisilvio v0.10.0 WHEELS 3.3, 3.4 makefile for WHEELS v0.10.0 history ALL Separate py2 and py3 packages to avoid installation errors. Fix #151 Running travis and tox with python 3.4.
This commit is contained in:
James Douglass
@@ -27,3 +27,4 @@ Patches and Suggestions
|
||||
- Jakub Janoszek
|
||||
- Marc Abramowitz
|
||||
- Alex Gaynor
|
||||
- James Douglass
|
||||
|
||||
+9
-1
@@ -24,6 +24,7 @@ Output formats supported:
|
||||
- HTML (Sets)
|
||||
- TSV (Sets)
|
||||
- CSV (Sets)
|
||||
- DBF (Sets)
|
||||
|
||||
Note that tablib *purposefully* excludes XML support. It always will. (Note: This is a joke. Pull requests are welcome.)
|
||||
|
||||
@@ -31,7 +32,7 @@ Overview
|
||||
--------
|
||||
|
||||
`tablib.Dataset()`
|
||||
A Dataset is a table of tabular data. It may or may not have a header row. They can be build and manipulated as raw Python datatypes (Lists of tuples|dictionaries). Datasets can be imported from JSON, YAML, and CSV; they can be exported to XLSX, XLS, ODS, JSON, YAML, CSV, TSV, and HTML.
|
||||
A Dataset is a table of tabular data. It may or may not have a header row. They can be build and manipulated as raw Python datatypes (Lists of tuples|dictionaries). Datasets can be imported from JSON, YAML, DBF, and CSV; they can be exported to XLSX, XLS, ODS, JSON, YAML, DBF, CSV, TSV, and HTML.
|
||||
|
||||
`tablib.Databook()`
|
||||
A Databook is a set of Datasets. The most common form of a Databook is an Excel file with multiple spreadsheets. Databooks can be imported from JSON and YAML; they can be exported to XLSX, XLS, ODS, JSON, and YAML.
|
||||
@@ -123,6 +124,13 @@ EXCEL!
|
||||
>>> with open('people.xls', 'wb') as f:
|
||||
... f.write(data.xls)
|
||||
|
||||
DBF!
|
||||
++++
|
||||
::
|
||||
|
||||
>>> with open('people.dbf', 'wb') as f:
|
||||
... f.write(data.dbf)
|
||||
|
||||
It's that easy.
|
||||
|
||||
|
||||
|
||||
@@ -50,6 +50,7 @@ if sys.version_info[0] == 2:
|
||||
'tablib.packages.openpyxl.reader',
|
||||
'tablib.packages.openpyxl.writer',
|
||||
'tablib.packages.yaml',
|
||||
'tablib.packages.dbfpy'
|
||||
])
|
||||
else:
|
||||
packages.extend([
|
||||
@@ -61,6 +62,7 @@ else:
|
||||
'tablib.packages.openpyxl3.reader',
|
||||
'tablib.packages.openpyxl3.writer',
|
||||
'tablib.packages.yaml3',
|
||||
'tablib.packages.dbfpy3'
|
||||
])
|
||||
|
||||
|
||||
|
||||
@@ -28,6 +28,7 @@ if is_py3:
|
||||
from tablib.packages import markup3 as markup
|
||||
from tablib.packages import openpyxl3 as openpyxl
|
||||
from tablib.packages.odf3 import opendocument, style, text, table
|
||||
import tablib.packages.dbfpy3 as dbfpy
|
||||
|
||||
import csv
|
||||
from io import StringIO
|
||||
@@ -49,5 +50,6 @@ else:
|
||||
from tablib.packages.odf import opendocument, style, text, table
|
||||
|
||||
from tablib.packages import unicodecsv as csv
|
||||
import tablib.packages.dbfpy as dbfpy
|
||||
|
||||
unicode = unicode
|
||||
|
||||
@@ -559,6 +559,29 @@ class Dataset(object):
|
||||
"""
|
||||
pass
|
||||
|
||||
@property
|
||||
def dbf():
|
||||
"""A dBASE representation of the :class:`Dataset` object.
|
||||
|
||||
A dataset object can also be imported by setting the :class:`Dataset.dbf` attribute:::
|
||||
|
||||
# To import data from an existing DBF file:
|
||||
data = tablib.Dataset()
|
||||
data.dbf = open('existing_table.dbf').read()
|
||||
|
||||
# to import data from an ASCII-encoded bytestring:
|
||||
data = tablib.Dataset()
|
||||
data.dbf = <bytestring of tabular data>
|
||||
|
||||
.. admonition:: Binary Warning
|
||||
|
||||
:class:`Dataset.dbf` contains binary data, so make sure to write in binary mode::
|
||||
|
||||
with open('output.dbf', 'wb') as f:
|
||||
f.write(data.dbf)
|
||||
"""
|
||||
pass
|
||||
|
||||
|
||||
# ----
|
||||
# Rows
|
||||
|
||||
@@ -11,5 +11,6 @@ from . import _tsv as tsv
|
||||
from . import _html as html
|
||||
from . import _xlsx as xlsx
|
||||
from . import _ods as ods
|
||||
from . import _dbf as dbf
|
||||
|
||||
available = (json, xls, yaml, csv, tsv, html, xlsx, ods)
|
||||
available = (json, xls, yaml, csv, dbf, tsv, html, xlsx, ods)
|
||||
|
||||
@@ -0,0 +1,93 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
""" Tablib - DBF Support.
|
||||
"""
|
||||
import tempfile
|
||||
import struct
|
||||
import os
|
||||
|
||||
from tablib.compat import StringIO
|
||||
from tablib.compat import dbfpy
|
||||
from tablib.compat import is_py3
|
||||
|
||||
if is_py3:
|
||||
from tablib.packages.dbfpy3 import dbf
|
||||
from tablib.packages.dbfpy3 import dbfnew
|
||||
from tablib.packages.dbfpy3 import record as dbfrecord
|
||||
import io
|
||||
else:
|
||||
from tablib.packages.dbfpy import dbf
|
||||
from tablib.packages.dbfpy import dbfnew
|
||||
from tablib.packages.dbfpy import record as dbfrecord
|
||||
|
||||
|
||||
title = 'dbf'
|
||||
extensions = ('csv',)
|
||||
|
||||
DEFAULT_ENCODING = 'utf-8'
|
||||
|
||||
def export_set(dataset):
|
||||
"""Returns DBF representation of a Dataset"""
|
||||
new_dbf = dbfnew.dbf_new()
|
||||
temp_file, temp_uri = tempfile.mkstemp()
|
||||
|
||||
# create the appropriate fields based on the contents of the first row
|
||||
first_row = dataset[0]
|
||||
for fieldname, field_value in zip(dataset.headers, first_row):
|
||||
if type(field_value) in [int, float]:
|
||||
new_dbf.add_field(fieldname, 'N', 10, 8)
|
||||
else:
|
||||
new_dbf.add_field(fieldname, 'C', 80)
|
||||
|
||||
new_dbf.write(temp_uri)
|
||||
|
||||
dbf_file = dbf.Dbf(temp_uri, readOnly=0)
|
||||
for row in dataset:
|
||||
record = dbfrecord.DbfRecord(dbf_file)
|
||||
for fieldname, field_value in zip(dataset.headers, row):
|
||||
record[fieldname] = field_value
|
||||
record.store()
|
||||
|
||||
dbf_file.close()
|
||||
dbf_stream = open(temp_uri, 'rb')
|
||||
if is_py3:
|
||||
stream = io.BytesIO(dbf_stream.read())
|
||||
else:
|
||||
stream = StringIO(dbf_stream.read())
|
||||
dbf_stream.close()
|
||||
os.remove(temp_uri)
|
||||
return stream.getvalue()
|
||||
|
||||
def import_set(dset, in_stream, headers=True):
|
||||
"""Returns a dataset from a DBF stream."""
|
||||
|
||||
dset.wipe()
|
||||
if is_py3:
|
||||
_dbf = dbf.Dbf(io.BytesIO(in_stream))
|
||||
else:
|
||||
_dbf = dbf.Dbf(StringIO(in_stream))
|
||||
dset.headers = _dbf.fieldNames
|
||||
for record in range(_dbf.recordCount):
|
||||
row = [_dbf[record][f] for f in _dbf.fieldNames]
|
||||
dset.append(row)
|
||||
|
||||
def detect(stream):
|
||||
"""Returns True if the given stream is valid DBF"""
|
||||
#_dbf = dbf.Table(StringIO(stream))
|
||||
try:
|
||||
if is_py3:
|
||||
if type(stream) is not bytes:
|
||||
stream = bytes(stream, 'utf-8')
|
||||
_dbf = dbf.Dbf(io.BytesIO(stream), readOnly=True)
|
||||
else:
|
||||
_dbf = dbf.Dbf(StringIO(stream), readOnly=True)
|
||||
return True
|
||||
except (ValueError, struct.error):
|
||||
# When we try to open up a file that's not a DBF, dbfpy raises a
|
||||
# ValueError.
|
||||
# When unpacking a string argument with less than 8 chars, struct.error is
|
||||
# raised.
|
||||
return False
|
||||
|
||||
|
||||
|
||||
@@ -0,0 +1,292 @@
|
||||
#! /usr/bin/env python
|
||||
"""DBF accessing helpers.
|
||||
|
||||
FIXME: more documentation needed
|
||||
|
||||
Examples:
|
||||
|
||||
Create new table, setup structure, add records:
|
||||
|
||||
dbf = Dbf(filename, new=True)
|
||||
dbf.addField(
|
||||
("NAME", "C", 15),
|
||||
("SURNAME", "C", 25),
|
||||
("INITIALS", "C", 10),
|
||||
("BIRTHDATE", "D"),
|
||||
)
|
||||
for (n, s, i, b) in (
|
||||
("John", "Miller", "YC", (1980, 10, 11)),
|
||||
("Andy", "Larkin", "", (1980, 4, 11)),
|
||||
):
|
||||
rec = dbf.newRecord()
|
||||
rec["NAME"] = n
|
||||
rec["SURNAME"] = s
|
||||
rec["INITIALS"] = i
|
||||
rec["BIRTHDATE"] = b
|
||||
rec.store()
|
||||
dbf.close()
|
||||
|
||||
Open existed dbf, read some data:
|
||||
|
||||
dbf = Dbf(filename, True)
|
||||
for rec in dbf:
|
||||
for fldName in dbf.fieldNames:
|
||||
print '%s:\t %s (%s)' % (fldName, rec[fldName],
|
||||
type(rec[fldName]))
|
||||
print
|
||||
dbf.close()
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] export INVALID_VALUE;
|
||||
Dbf: added .ignoreErrors, .INVALID_VALUE
|
||||
04-jul-2006 [als] added export declaration
|
||||
20-dec-2005 [yc] removed fromStream and newDbf methods:
|
||||
use argument of __init__ call must be used instead;
|
||||
added class fields pointing to the header and
|
||||
record classes.
|
||||
17-dec-2005 [yc] split to several modules; reimplemented
|
||||
13-dec-2005 [yc] adapted to the changes of the `strutil` module.
|
||||
13-sep-2002 [als] support FoxPro Timestamp datatype
|
||||
15-nov-1999 [jjk] documentation updates, add demo
|
||||
24-aug-1998 [jjk] add some encodeValue methods (not tested), other tweaks
|
||||
08-jun-1998 [jjk] fix problems, add more features
|
||||
20-feb-1998 [jjk] fix problems, add more features
|
||||
19-feb-1998 [jjk] add create/write capabilities
|
||||
18-feb-1998 [jjk] from dbfload.py
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.7 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 09:23:13 $"[7:-2]
|
||||
__author__ = "Jeff Kunce <kuncej@mail.conservation.state.mo.us>"
|
||||
|
||||
__all__ = ["Dbf"]
|
||||
|
||||
from . import header
|
||||
from .import record
|
||||
from utils import INVALID_VALUE
|
||||
|
||||
class Dbf(object):
|
||||
"""DBF accessor.
|
||||
|
||||
FIXME:
|
||||
docs and examples needed (dont' forget to tell
|
||||
about problems adding new fields on the fly)
|
||||
|
||||
Implementation notes:
|
||||
``_new`` field is used to indicate whether this is
|
||||
a new data table. `addField` could be used only for
|
||||
the new tables! If at least one record was appended
|
||||
to the table it's structure couldn't be changed.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("name", "header", "stream",
|
||||
"_changed", "_new", "_ignore_errors")
|
||||
|
||||
HeaderClass = header.DbfHeader
|
||||
RecordClass = record.DbfRecord
|
||||
INVALID_VALUE = INVALID_VALUE
|
||||
|
||||
## initialization and creation helpers
|
||||
|
||||
def __init__(self, f, readOnly=False, new=False, ignoreErrors=False):
|
||||
"""Initialize instance.
|
||||
|
||||
Arguments:
|
||||
f:
|
||||
Filename or file-like object.
|
||||
new:
|
||||
True if new data table must be created. Assume
|
||||
data table exists if this argument is False.
|
||||
readOnly:
|
||||
if ``f`` argument is a string file will
|
||||
be opend in read-only mode; in other cases
|
||||
this argument is ignored. This argument is ignored
|
||||
even if ``new`` argument is True.
|
||||
headerObj:
|
||||
`header.DbfHeader` instance or None. If this argument
|
||||
is None, new empty header will be used with the
|
||||
all fields set by default.
|
||||
ignoreErrors:
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
"""
|
||||
if isinstance(f, basestring):
|
||||
# a filename
|
||||
self.name = f
|
||||
if new:
|
||||
# new table (table file must be
|
||||
# created or opened and truncated)
|
||||
self.stream = file(f, "w+b")
|
||||
else:
|
||||
# tabe file must exist
|
||||
self.stream = file(f, ("r+b", "rb")[bool(readOnly)])
|
||||
else:
|
||||
# a stream
|
||||
self.name = getattr(f, "name", "")
|
||||
self.stream = f
|
||||
if new:
|
||||
# if this is a new table, header will be empty
|
||||
self.header = self.HeaderClass()
|
||||
else:
|
||||
# or instantiated using stream
|
||||
self.header = self.HeaderClass.fromStream(self.stream)
|
||||
self.ignoreErrors = ignoreErrors
|
||||
self._new = bool(new)
|
||||
self._changed = False
|
||||
|
||||
## properties
|
||||
|
||||
closed = property(lambda self: self.stream.closed)
|
||||
recordCount = property(lambda self: self.header.recordCount)
|
||||
fieldNames = property(
|
||||
lambda self: [_fld.name for _fld in self.header.fields])
|
||||
fieldDefs = property(lambda self: self.header.fields)
|
||||
changed = property(lambda self: self._changed or self.header.changed)
|
||||
|
||||
def ignoreErrors(self, value):
|
||||
"""Update `ignoreErrors` flag on the header object and self"""
|
||||
self.header.ignoreErrors = self._ignore_errors = bool(value)
|
||||
ignoreErrors = property(
|
||||
lambda self: self._ignore_errors,
|
||||
ignoreErrors,
|
||||
doc="""Error processing mode for DBF field value conversion
|
||||
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
""")
|
||||
|
||||
## protected methods
|
||||
|
||||
def _fixIndex(self, index):
|
||||
"""Return fixed index.
|
||||
|
||||
This method fails if index isn't a numeric object
|
||||
(long or int). Or index isn't in a valid range
|
||||
(less or equal to the number of records in the db).
|
||||
|
||||
If ``index`` is a negative number, it will be
|
||||
treated as a negative indexes for list objects.
|
||||
|
||||
Return:
|
||||
Return value is numeric object maning valid index.
|
||||
|
||||
"""
|
||||
if not isinstance(index, (int, long)):
|
||||
raise TypeError("Index must be a numeric object")
|
||||
if index < 0:
|
||||
# index from the right side
|
||||
# fix it to the left-side index
|
||||
index += len(self) + 1
|
||||
if index >= len(self):
|
||||
raise IndexError("Record index out of range")
|
||||
return index
|
||||
|
||||
## iterface methods
|
||||
|
||||
def close(self):
|
||||
self.flush()
|
||||
self.stream.close()
|
||||
|
||||
def flush(self):
|
||||
"""Flush data to the associated stream."""
|
||||
if self.changed:
|
||||
self.header.setCurrentDate()
|
||||
self.header.write(self.stream)
|
||||
self.stream.flush()
|
||||
self._changed = False
|
||||
|
||||
def indexOfFieldName(self, name):
|
||||
"""Index of field named ``name``."""
|
||||
# FIXME: move this to header class
|
||||
return self.header.fields.index(name)
|
||||
|
||||
def newRecord(self):
|
||||
"""Return new record, which belong to this table."""
|
||||
return self.RecordClass(self)
|
||||
|
||||
def append(self, record):
|
||||
"""Append ``record`` to the database."""
|
||||
record.index = self.header.recordCount
|
||||
record._write()
|
||||
self.header.recordCount += 1
|
||||
self._changed = True
|
||||
self._new = False
|
||||
|
||||
def addField(self, *defs):
|
||||
"""Add field definitions.
|
||||
|
||||
For more information see `header.DbfHeader.addField`.
|
||||
|
||||
"""
|
||||
if self._new:
|
||||
self.header.addField(*defs)
|
||||
else:
|
||||
raise TypeError("At least one record was added, "
|
||||
"structure can't be changed")
|
||||
|
||||
## 'magic' methods (representation and sequence interface)
|
||||
|
||||
def __repr__(self):
|
||||
return "Dbf stream '%s'\n" % self.stream + repr(self.header)
|
||||
|
||||
def __len__(self):
|
||||
"""Return number of records."""
|
||||
return self.recordCount
|
||||
|
||||
def __getitem__(self, index):
|
||||
"""Return `DbfRecord` instance."""
|
||||
return self.RecordClass.fromStream(self, self._fixIndex(index))
|
||||
|
||||
def __setitem__(self, index, record):
|
||||
"""Write `DbfRecord` instance to the stream."""
|
||||
record.index = self._fixIndex(index)
|
||||
record._write()
|
||||
self._changed = True
|
||||
self._new = False
|
||||
|
||||
#def __del__(self):
|
||||
# """Flush stream upon deletion of the object."""
|
||||
# self.flush()
|
||||
|
||||
|
||||
def demoRead(filename):
|
||||
_dbf = Dbf(filename, True)
|
||||
for _rec in _dbf:
|
||||
print
|
||||
print(repr(_rec))
|
||||
_dbf.close()
|
||||
|
||||
def demoCreate(filename):
|
||||
_dbf = Dbf(filename, new=True)
|
||||
_dbf.addField(
|
||||
("NAME", "C", 15),
|
||||
("SURNAME", "C", 25),
|
||||
("INITIALS", "C", 10),
|
||||
("BIRTHDATE", "D"),
|
||||
)
|
||||
for (_n, _s, _i, _b) in (
|
||||
("John", "Miller", "YC", (1981, 1, 2)),
|
||||
("Andy", "Larkin", "AL", (1982, 3, 4)),
|
||||
("Bill", "Clinth", "", (1983, 5, 6)),
|
||||
("Bobb", "McNail", "", (1984, 7, 8)),
|
||||
):
|
||||
_rec = _dbf.newRecord()
|
||||
_rec["NAME"] = _n
|
||||
_rec["SURNAME"] = _s
|
||||
_rec["INITIALS"] = _i
|
||||
_rec["BIRTHDATE"] = _b
|
||||
_rec.store()
|
||||
print(repr(_dbf))
|
||||
_dbf.close()
|
||||
|
||||
if (__name__=='__main__'):
|
||||
import sys
|
||||
_name = len(sys.argv) > 1 and sys.argv[1] or "county.dbf"
|
||||
demoCreate(_name)
|
||||
demoRead(_name)
|
||||
|
||||
# vim: set et sw=4 sts=4 :
|
||||
@@ -0,0 +1,188 @@
|
||||
#!/usr/bin/python
|
||||
""".DBF creation helpers.
|
||||
|
||||
Note: this is a legacy interface. New code should use Dbf class
|
||||
for table creation (see examples in dbf.py)
|
||||
|
||||
TODO:
|
||||
- handle Memo fields.
|
||||
- check length of the fields accoring to the
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
|
||||
"""
|
||||
"""History (most recent first)
|
||||
04-jul-2006 [als] added export declaration;
|
||||
updated for dbfpy 2.0
|
||||
15-dec-2005 [yc] define dbf_new.__slots__
|
||||
14-dec-2005 [yc] added vim modeline; retab'd; added doc-strings;
|
||||
dbf_new now is a new class (inherited from object)
|
||||
??-jun-2000 [--] added by Hans Fiby
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.4 $"[11:-2]
|
||||
__date__ = "$Date: 2006/07/04 08:18:18 $"[7:-2]
|
||||
|
||||
__all__ = ["dbf_new"]
|
||||
|
||||
from dbf import *
|
||||
from fields import *
|
||||
from header import *
|
||||
from record import *
|
||||
|
||||
class _FieldDefinition(object):
|
||||
"""Field definition.
|
||||
|
||||
This is a simple structure, which contains ``name``, ``type``,
|
||||
``len``, ``dec`` and ``cls`` fields.
|
||||
|
||||
Objects also implement get/setitem magic functions, so fields
|
||||
could be accessed via sequence iterface, where 'name' has
|
||||
index 0, 'type' index 1, 'len' index 2, 'dec' index 3 and
|
||||
'cls' could be located at index 4.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = "name", "type", "len", "dec", "cls"
|
||||
|
||||
# WARNING: be attentive - dictionaries are mutable!
|
||||
FLD_TYPES = {
|
||||
# type: (cls, len)
|
||||
"C": (DbfCharacterFieldDef, None),
|
||||
"N": (DbfNumericFieldDef, None),
|
||||
"L": (DbfLogicalFieldDef, 1),
|
||||
# FIXME: support memos
|
||||
# "M": (DbfMemoFieldDef),
|
||||
"D": (DbfDateFieldDef, 8),
|
||||
# FIXME: I'm not sure length should be 14 characters!
|
||||
# but temporary I use it, cuz date is 8 characters
|
||||
# and time 6 (hhmmss)
|
||||
"T": (DbfDateTimeFieldDef, 14),
|
||||
}
|
||||
|
||||
def __init__(self, name, type, len=None, dec=0):
|
||||
_cls, _len = self.FLD_TYPES[type]
|
||||
if _len is None:
|
||||
if len is None:
|
||||
raise ValueError("Field length must be defined")
|
||||
_len = len
|
||||
self.name = name
|
||||
self.type = type
|
||||
self.len = _len
|
||||
self.dec = dec
|
||||
self.cls = _cls
|
||||
|
||||
def getDbfField(self):
|
||||
"Return `DbfFieldDef` instance from the current definition."
|
||||
return self.cls(self.name, self.len, self.dec)
|
||||
|
||||
def appendToHeader(self, dbfh):
|
||||
"""Create a `DbfFieldDef` instance and append it to the dbf header.
|
||||
|
||||
Arguments:
|
||||
dbfh: `DbfHeader` instance.
|
||||
|
||||
"""
|
||||
_dbff = self.getDbfField()
|
||||
dbfh.addField(_dbff)
|
||||
|
||||
|
||||
class dbf_new(object):
|
||||
"""New .DBF creation helper.
|
||||
|
||||
Example Usage:
|
||||
|
||||
dbfn = dbf_new()
|
||||
dbfn.add_field("name",'C',80)
|
||||
dbfn.add_field("price",'N',10,2)
|
||||
dbfn.add_field("date",'D',8)
|
||||
dbfn.write("tst.dbf")
|
||||
|
||||
Note:
|
||||
This module cannot handle Memo-fields,
|
||||
they are special.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("fields",)
|
||||
|
||||
FieldDefinitionClass = _FieldDefinition
|
||||
|
||||
def __init__(self):
|
||||
self.fields = []
|
||||
|
||||
def add_field(self, name, typ, len, dec=0):
|
||||
"""Add field definition.
|
||||
|
||||
Arguments:
|
||||
name:
|
||||
field name (str object). field name must not
|
||||
contain ASCII NULs and it's length shouldn't
|
||||
exceed 10 characters.
|
||||
typ:
|
||||
type of the field. this must be a single character
|
||||
from the "CNLMDT" set meaning character, numeric,
|
||||
logical, memo, date and date/time respectively.
|
||||
len:
|
||||
length of the field. this argument is used only for
|
||||
the character and numeric fields. all other fields
|
||||
have fixed length.
|
||||
FIXME: use None as a default for this argument?
|
||||
dec:
|
||||
decimal precision. used only for the numric fields.
|
||||
|
||||
"""
|
||||
self.fields.append(self.FieldDefinitionClass(name, typ, len, dec))
|
||||
|
||||
def write(self, filename):
|
||||
"""Create empty .DBF file using current structure."""
|
||||
_dbfh = DbfHeader()
|
||||
_dbfh.setCurrentDate()
|
||||
for _fldDef in self.fields:
|
||||
_fldDef.appendToHeader(_dbfh)
|
||||
_dbfStream = file(filename, "wb")
|
||||
_dbfh.write(_dbfStream)
|
||||
_dbfStream.close()
|
||||
|
||||
def write_stream(self, stream):
|
||||
_dbfh = DbfHeader()
|
||||
_dbfh.setCurrentDate()
|
||||
for _fldDef in self.fields:
|
||||
_fldDef.appendToHeader(_dbfh)
|
||||
_dbfh.write(stream)
|
||||
|
||||
|
||||
if (__name__=='__main__'):
|
||||
# create a new DBF-File
|
||||
dbfn=dbf_new()
|
||||
dbfn.add_field("name",'C',80)
|
||||
dbfn.add_field("price",'N',10,2)
|
||||
dbfn.add_field("date",'D',8)
|
||||
dbfn.write("tst.dbf")
|
||||
# test new dbf
|
||||
print "*** created tst.dbf: ***"
|
||||
dbft = Dbf('tst.dbf', readOnly=0)
|
||||
print repr(dbft)
|
||||
# add a record
|
||||
rec=DbfRecord(dbft)
|
||||
rec['name']='something'
|
||||
rec['price']=10.5
|
||||
rec['date']=(2000,1,12)
|
||||
rec.store()
|
||||
# add another record
|
||||
rec=DbfRecord(dbft)
|
||||
rec['name']='foo and bar'
|
||||
rec['price']=12234
|
||||
rec['date']=(1992,7,15)
|
||||
rec.store()
|
||||
|
||||
# show the records
|
||||
print "*** inserted 2 records into tst.dbf: ***"
|
||||
print repr(dbft)
|
||||
for i1 in range(len(dbft)):
|
||||
rec = dbft[i1]
|
||||
for fldName in dbft.fieldNames:
|
||||
print '%s:\t %s'%(fldName, rec[fldName])
|
||||
print
|
||||
dbft.close()
|
||||
|
||||
# vim: set et sts=4 sw=4 :
|
||||
@@ -0,0 +1,466 @@
|
||||
"""DBF fields definitions.
|
||||
|
||||
TODO:
|
||||
- make memos work
|
||||
"""
|
||||
"""History (most recent first):
|
||||
26-may-2009 [als] DbfNumericFieldDef.decodeValue: strip zero bytes
|
||||
05-feb-2009 [als] DbfDateFieldDef.encodeValue: empty arg produces empty date
|
||||
16-sep-2008 [als] DbfNumericFieldDef decoding looks for decimal point
|
||||
in the value to select float or integer return type
|
||||
13-mar-2008 [als] check field name length in constructor
|
||||
11-feb-2007 [als] handle value conversion errors
|
||||
10-feb-2007 [als] DbfFieldDef: added .rawFromRecord()
|
||||
01-dec-2006 [als] Timestamp columns use None for empty values
|
||||
31-oct-2006 [als] support field types 'F' (float), 'I' (integer)
|
||||
and 'Y' (currency);
|
||||
automate export and registration of field classes
|
||||
04-jul-2006 [als] added export declaration
|
||||
10-mar-2006 [als] decode empty values for Date and Logical fields;
|
||||
show field name in errors
|
||||
10-mar-2006 [als] fix Numeric value decoding: according to spec,
|
||||
value always is string representation of the number;
|
||||
ensure that encoded Numeric value fits into the field
|
||||
20-dec-2005 [yc] use field names in upper case
|
||||
15-dec-2005 [yc] field definitions moved from `dbf`.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.14 $"[11:-2]
|
||||
__date__ = "$Date: 2009/05/26 05:16:51 $"[7:-2]
|
||||
|
||||
__all__ = ["lookupFor",] # field classes added at the end of the module
|
||||
|
||||
import datetime
|
||||
import struct
|
||||
import sys
|
||||
|
||||
from . import utils
|
||||
|
||||
## abstract definitions
|
||||
|
||||
class DbfFieldDef(object):
|
||||
"""Abstract field definition.
|
||||
|
||||
Child classes must override ``type`` class attribute to provide datatype
|
||||
infromation of the field definition. For more info about types visit
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
|
||||
Also child classes must override ``defaultValue`` field to provide
|
||||
default value for the field value.
|
||||
|
||||
If child class has fixed length ``length`` class attribute must be
|
||||
overriden and set to the valid value. None value means, that field
|
||||
isn't of fixed length.
|
||||
|
||||
Note: ``name`` field must not be changed after instantiation.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("name", "length", "decimalCount",
|
||||
"start", "end", "ignoreErrors")
|
||||
|
||||
# length of the field, None in case of variable-length field,
|
||||
# or a number if this field is a fixed-length field
|
||||
length = None
|
||||
|
||||
# field type. for more information about fields types visit
|
||||
# `http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
# must be overriden in child classes
|
||||
typeCode = None
|
||||
|
||||
# default value for the field. this field must be
|
||||
# overriden in child classes
|
||||
defaultValue = None
|
||||
|
||||
def __init__(self, name, length=None, decimalCount=None,
|
||||
start=None, stop=None, ignoreErrors=False,
|
||||
):
|
||||
"""Initialize instance."""
|
||||
assert self.typeCode is not None, "Type code must be overriden"
|
||||
assert self.defaultValue is not None, "Default value must be overriden"
|
||||
## fix arguments
|
||||
if len(name) >10:
|
||||
raise ValueError("Field name \"%s\" is too long" % name)
|
||||
name = str(name).upper()
|
||||
if self.__class__.length is None:
|
||||
if length is None:
|
||||
raise ValueError("[%s] Length isn't specified" % name)
|
||||
length = int(length)
|
||||
if length <= 0:
|
||||
raise ValueError("[%s] Length must be a positive integer"
|
||||
% name)
|
||||
else:
|
||||
length = self.length
|
||||
if decimalCount is None:
|
||||
decimalCount = 0
|
||||
## set fields
|
||||
self.name = name
|
||||
# FIXME: validate length according to the specification at
|
||||
# http://www.clicketyclick.dk/databases/xbase/format/data_types.html
|
||||
self.length = length
|
||||
self.decimalCount = decimalCount
|
||||
self.ignoreErrors = ignoreErrors
|
||||
self.start = start
|
||||
self.end = stop
|
||||
|
||||
def __cmp__(self, other):
|
||||
return cmp(self.name, str(other).upper())
|
||||
|
||||
def __hash__(self):
|
||||
return hash(self.name)
|
||||
|
||||
def fromString(cls, string, start, ignoreErrors=False):
|
||||
"""Decode dbf field definition from the string data.
|
||||
|
||||
Arguments:
|
||||
string:
|
||||
a string, dbf definition is decoded from. length of
|
||||
the string must be 32 bytes.
|
||||
start:
|
||||
position in the database file.
|
||||
ignoreErrors:
|
||||
initial error processing mode for the new field (boolean)
|
||||
|
||||
"""
|
||||
assert len(string) == 32
|
||||
_length = ord(string[16])
|
||||
return cls(utils.unzfill(string)[:11], _length, ord(string[17]),
|
||||
start, start + _length, ignoreErrors=ignoreErrors)
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
def toString(self):
|
||||
"""Return encoded field definition.
|
||||
|
||||
Return:
|
||||
Return value is a string object containing encoded
|
||||
definition of this field.
|
||||
|
||||
"""
|
||||
if sys.version_info < (2, 4):
|
||||
# earlier versions did not support padding character
|
||||
_name = self.name[:11] + "\0" * (11 - len(self.name))
|
||||
else:
|
||||
_name = self.name.ljust(11, '\0')
|
||||
return (
|
||||
_name +
|
||||
self.typeCode +
|
||||
#data address
|
||||
chr(0) * 4 +
|
||||
chr(self.length) +
|
||||
chr(self.decimalCount) +
|
||||
chr(0) * 14
|
||||
)
|
||||
|
||||
def __repr__(self):
|
||||
return "%-10s %1s %3d %3d" % self.fieldInfo()
|
||||
|
||||
def fieldInfo(self):
|
||||
"""Return field information.
|
||||
|
||||
Return:
|
||||
Return value is a (name, type, length, decimals) tuple.
|
||||
|
||||
"""
|
||||
return (self.name, self.typeCode, self.length, self.decimalCount)
|
||||
|
||||
def rawFromRecord(self, record):
|
||||
"""Return a "raw" field value from the record string."""
|
||||
return record[self.start:self.end]
|
||||
|
||||
def decodeFromRecord(self, record):
|
||||
"""Return decoded field value from the record string."""
|
||||
try:
|
||||
return self.decodeValue(self.rawFromRecord(record))
|
||||
except:
|
||||
if self.ignoreErrors:
|
||||
return utils.INVALID_VALUE
|
||||
else:
|
||||
raise
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return decoded value from string value.
|
||||
|
||||
This method shouldn't be used publicly. It's called from the
|
||||
`decodeFromRecord` method.
|
||||
|
||||
This is an abstract method and it must be overridden in child classes.
|
||||
"""
|
||||
raise NotImplementedError
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return str object containing encoded field value.
|
||||
|
||||
This is an abstract method and it must be overriden in child classes.
|
||||
"""
|
||||
raise NotImplementedError
|
||||
|
||||
## real classes
|
||||
|
||||
class DbfCharacterFieldDef(DbfFieldDef):
|
||||
"""Definition of the character field."""
|
||||
|
||||
typeCode = "C"
|
||||
defaultValue = ""
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return string object.
|
||||
|
||||
Return value is a ``value`` argument with stripped right spaces.
|
||||
|
||||
"""
|
||||
return value.rstrip(" ")
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return raw data string encoded from a ``value``."""
|
||||
return str(value)[:self.length].ljust(self.length)
|
||||
|
||||
|
||||
class DbfNumericFieldDef(DbfFieldDef):
|
||||
"""Definition of the numeric field."""
|
||||
|
||||
typeCode = "N"
|
||||
# XXX: now I'm not sure it was a good idea to make a class field
|
||||
# `defaultValue` instead of a generic method as it was implemented
|
||||
# previously -- it's ok with all types except number, cuz
|
||||
# if self.decimalCount is 0, we should return 0 and 0.0 otherwise.
|
||||
defaultValue = 0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a number decoded from ``value``.
|
||||
|
||||
If decimals is zero, value will be decoded as an integer;
|
||||
or as a float otherwise.
|
||||
|
||||
Return:
|
||||
Return value is a int (long) or float instance.
|
||||
|
||||
"""
|
||||
value = value.strip(" \0")
|
||||
if "." in value:
|
||||
# a float (has decimal separator)
|
||||
return float(value)
|
||||
elif value:
|
||||
# must be an integer
|
||||
return int(value)
|
||||
else:
|
||||
return 0
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
_rv = ("%*.*f" % (self.length, self.decimalCount, value))
|
||||
if len(_rv) > self.length:
|
||||
_ppos = _rv.find(".")
|
||||
if 0 <= _ppos <= self.length:
|
||||
_rv = _rv[:self.length]
|
||||
else:
|
||||
raise ValueError("[%s] Numeric overflow: %s (field width: %i)"
|
||||
% (self.name, _rv, self.length))
|
||||
return _rv
|
||||
|
||||
class DbfFloatFieldDef(DbfNumericFieldDef):
|
||||
"""Definition of the float field - same as numeric."""
|
||||
|
||||
typeCode = "F"
|
||||
|
||||
class DbfIntegerFieldDef(DbfFieldDef):
|
||||
"""Definition of the integer field."""
|
||||
|
||||
typeCode = "I"
|
||||
length = 4
|
||||
defaultValue = 0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return an integer number decoded from ``value``."""
|
||||
return struct.unpack("<i", value)[0]
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
return struct.pack("<i", int(value))
|
||||
|
||||
class DbfCurrencyFieldDef(DbfFieldDef):
|
||||
"""Definition of the currency field."""
|
||||
|
||||
typeCode = "Y"
|
||||
length = 8
|
||||
defaultValue = 0.0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return float number decoded from ``value``."""
|
||||
return struct.unpack("<q", value)[0] / 10000.
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
return struct.pack("<q", round(value * 10000))
|
||||
|
||||
class DbfLogicalFieldDef(DbfFieldDef):
|
||||
"""Definition of the logical field."""
|
||||
|
||||
typeCode = "L"
|
||||
defaultValue = -1
|
||||
length = 1
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return True, False or -1 decoded from ``value``."""
|
||||
# Note: value always is 1-char string
|
||||
if value == "?":
|
||||
return -1
|
||||
if value in "NnFf ":
|
||||
return False
|
||||
if value in "YyTt":
|
||||
return True
|
||||
raise ValueError("[%s] Invalid logical value %r" % (self.name, value))
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a character from the "TF?" set.
|
||||
|
||||
Return:
|
||||
Return value is "T" if ``value`` is True
|
||||
"?" if value is -1 or False otherwise.
|
||||
|
||||
"""
|
||||
if value is True:
|
||||
return "T"
|
||||
if value == -1:
|
||||
return "?"
|
||||
return "F"
|
||||
|
||||
|
||||
class DbfMemoFieldDef(DbfFieldDef):
|
||||
"""Definition of the memo field.
|
||||
|
||||
Note: memos aren't currenly completely supported.
|
||||
|
||||
"""
|
||||
|
||||
typeCode = "M"
|
||||
defaultValue = " " * 10
|
||||
length = 10
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return int .dbt block number decoded from the string object."""
|
||||
#return int(value)
|
||||
raise NotImplementedError
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return raw data string encoded from a ``value``.
|
||||
|
||||
Note: this is an internal method.
|
||||
|
||||
"""
|
||||
#return str(value)[:self.length].ljust(self.length)
|
||||
raise NotImplementedError
|
||||
|
||||
|
||||
class DbfDateFieldDef(DbfFieldDef):
|
||||
"""Definition of the date field."""
|
||||
|
||||
typeCode = "D"
|
||||
defaultValue = utils.classproperty(lambda cls: datetime.date.today())
|
||||
# "yyyymmdd" gives us 8 characters
|
||||
length = 8
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a ``datetime.date`` instance decoded from ``value``."""
|
||||
if value.strip():
|
||||
return utils.getDate(value)
|
||||
else:
|
||||
return None
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a string-encoded value.
|
||||
|
||||
``value`` argument should be a value suitable for the
|
||||
`utils.getDate` call.
|
||||
|
||||
Return:
|
||||
Return value is a string in format "yyyymmdd".
|
||||
|
||||
"""
|
||||
if value:
|
||||
return utils.getDate(value).strftime("%Y%m%d")
|
||||
else:
|
||||
return " " * self.length
|
||||
|
||||
|
||||
class DbfDateTimeFieldDef(DbfFieldDef):
|
||||
"""Definition of the timestamp field."""
|
||||
|
||||
# a difference between JDN (Julian Day Number)
|
||||
# and GDN (Gregorian Day Number). note, that GDN < JDN
|
||||
JDN_GDN_DIFF = 1721425
|
||||
typeCode = "T"
|
||||
defaultValue = utils.classproperty(lambda cls: datetime.datetime.now())
|
||||
# two 32-bits integers representing JDN and amount of
|
||||
# milliseconds respectively gives us 8 bytes.
|
||||
# note, that values must be encoded in LE byteorder.
|
||||
length = 8
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a `datetime.datetime` instance."""
|
||||
assert len(value) == self.length
|
||||
# LE byteorder
|
||||
_jdn, _msecs = struct.unpack("<2I", value)
|
||||
if _jdn >= 1:
|
||||
_rv = datetime.datetime.fromordinal(_jdn - self.JDN_GDN_DIFF)
|
||||
_rv += datetime.timedelta(0, _msecs / 1000.0)
|
||||
else:
|
||||
# empty date
|
||||
_rv = None
|
||||
return _rv
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a string-encoded ``value``."""
|
||||
if value:
|
||||
value = utils.getDateTime(value)
|
||||
# LE byteorder
|
||||
_rv = struct.pack("<2I", value.toordinal() + self.JDN_GDN_DIFF,
|
||||
(value.hour * 3600 + value.minute * 60 + value.second) * 1000)
|
||||
else:
|
||||
_rv = "\0" * self.length
|
||||
assert len(_rv) == self.length
|
||||
return _rv
|
||||
|
||||
|
||||
_fieldsRegistry = {}
|
||||
|
||||
def registerField(fieldCls):
|
||||
"""Register field definition class.
|
||||
|
||||
``fieldCls`` should be subclass of the `DbfFieldDef`.
|
||||
|
||||
Use `lookupFor` to retrieve field definition class
|
||||
by the type code.
|
||||
|
||||
"""
|
||||
assert fieldCls.typeCode is not None, "Type code isn't defined"
|
||||
# XXX: use fieldCls.typeCode.upper()? in case of any decign
|
||||
# don't forget to look to the same comment in ``lookupFor`` method
|
||||
_fieldsRegistry[fieldCls.typeCode] = fieldCls
|
||||
|
||||
|
||||
def lookupFor(typeCode):
|
||||
"""Return field definition class for the given type code.
|
||||
|
||||
``typeCode`` must be a single character. That type should be
|
||||
previously registered.
|
||||
|
||||
Use `registerField` to register new field class.
|
||||
|
||||
Return:
|
||||
Return value is a subclass of the `DbfFieldDef`.
|
||||
|
||||
"""
|
||||
# XXX: use typeCode.upper()? in case of any decign don't
|
||||
# forget to look to the same comment in ``registerField``
|
||||
return _fieldsRegistry[typeCode]
|
||||
|
||||
## register generic types
|
||||
|
||||
for (_name, _val) in globals().items():
|
||||
if isinstance(_val, type) and issubclass(_val, DbfFieldDef) \
|
||||
and (_name != "DbfFieldDef"):
|
||||
__all__.append(_name)
|
||||
registerField(_val)
|
||||
del _name, _val
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,275 @@
|
||||
"""DBF header definition.
|
||||
|
||||
TODO:
|
||||
- handle encoding of the character fields
|
||||
(encoding information stored in the DBF header)
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
16-sep-2010 [als] fromStream: fix century of the last update field
|
||||
11-feb-2007 [als] added .ignoreErrors
|
||||
10-feb-2007 [als] added __getitem__: return field definitions
|
||||
by field name or field number (zero-based)
|
||||
04-jul-2006 [als] added export declaration
|
||||
15-dec-2005 [yc] created
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.6 $"[11:-2]
|
||||
__date__ = "$Date: 2010/09/16 05:06:39 $"[7:-2]
|
||||
|
||||
__all__ = ["DbfHeader"]
|
||||
|
||||
try:
|
||||
import cStringIO
|
||||
except ImportError:
|
||||
# when we're in python3, we cStringIO has been replaced by io.StringIO
|
||||
import io as cStringIO
|
||||
import datetime
|
||||
import struct
|
||||
import time
|
||||
|
||||
from . import fields
|
||||
from . import utils
|
||||
|
||||
|
||||
class DbfHeader(object):
|
||||
"""Dbf header definition.
|
||||
|
||||
For more information about dbf header format visit
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/dbf.html#DBF_STRUCT`
|
||||
|
||||
Examples:
|
||||
Create an empty dbf header and add some field definitions:
|
||||
dbfh = DbfHeader()
|
||||
dbfh.addField(("name", "C", 10))
|
||||
dbfh.addField(("date", "D"))
|
||||
dbfh.addField(DbfNumericFieldDef("price", 5, 2))
|
||||
Create a dbf header with field definitions:
|
||||
dbfh = DbfHeader([
|
||||
("name", "C", 10),
|
||||
("date", "D"),
|
||||
DbfNumericFieldDef("price", 5, 2),
|
||||
])
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("signature", "fields", "lastUpdate", "recordLength",
|
||||
"recordCount", "headerLength", "changed", "_ignore_errors")
|
||||
|
||||
## instance construction and initialization methods
|
||||
|
||||
def __init__(self, fields=None, headerLength=0, recordLength=0,
|
||||
recordCount=0, signature=0x03, lastUpdate=None, ignoreErrors=False,
|
||||
):
|
||||
"""Initialize instance.
|
||||
|
||||
Arguments:
|
||||
fields:
|
||||
a list of field definitions;
|
||||
recordLength:
|
||||
size of the records;
|
||||
headerLength:
|
||||
size of the header;
|
||||
recordCount:
|
||||
number of records stored in DBF;
|
||||
signature:
|
||||
version number (aka signature). using 0x03 as a default meaning
|
||||
"File without DBT". for more information about this field visit
|
||||
``http://www.clicketyclick.dk/databases/xbase/format/dbf.html#DBF_NOTE_1_TARGET``
|
||||
lastUpdate:
|
||||
date of the DBF's update. this could be a string ('yymmdd' or
|
||||
'yyyymmdd'), timestamp (int or float), datetime/date value,
|
||||
a sequence (assuming (yyyy, mm, dd, ...)) or an object having
|
||||
callable ``ticks`` field.
|
||||
ignoreErrors:
|
||||
error processing mode for DBF fields (boolean)
|
||||
|
||||
"""
|
||||
self.signature = signature
|
||||
if fields is None:
|
||||
self.fields = []
|
||||
else:
|
||||
self.fields = list(fields)
|
||||
self.lastUpdate = utils.getDate(lastUpdate)
|
||||
self.recordLength = recordLength
|
||||
self.headerLength = headerLength
|
||||
self.recordCount = recordCount
|
||||
self.ignoreErrors = ignoreErrors
|
||||
# XXX: I'm not sure this is safe to
|
||||
# initialize `self.changed` in this way
|
||||
self.changed = bool(self.fields)
|
||||
|
||||
# @classmethod
|
||||
def fromString(cls, string):
|
||||
"""Return header instance from the string object."""
|
||||
return cls.fromStream(cStringIO.StringIO(str(string)))
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
# @classmethod
|
||||
def fromStream(cls, stream):
|
||||
"""Return header object from the stream."""
|
||||
stream.seek(0)
|
||||
_data = stream.read(32)
|
||||
(_cnt, _hdrLen, _recLen) = struct.unpack("<I2H", _data[4:12])
|
||||
#reserved = _data[12:32]
|
||||
_year = ord(_data[1])
|
||||
if _year < 80:
|
||||
# dBase II started at 1980. It is quite unlikely
|
||||
# that actual last update date is before that year.
|
||||
_year += 2000
|
||||
else:
|
||||
_year += 1900
|
||||
## create header object
|
||||
_obj = cls(None, _hdrLen, _recLen, _cnt, ord(_data[0]),
|
||||
(_year, ord(_data[2]), ord(_data[3])))
|
||||
## append field definitions
|
||||
# position 0 is for the deletion flag
|
||||
_pos = 1
|
||||
_data = stream.read(1)
|
||||
|
||||
# The field definitions are ended either by \x0D OR a newline
|
||||
# character, so we need to handle both when reading from a stream.
|
||||
# When writing, dbfpy appears to write newlines instead of \x0D.
|
||||
while _data[0] not in ["\x0D", "\n"]:
|
||||
_data += stream.read(31)
|
||||
_fld = fields.lookupFor(_data[11]).fromString(_data, _pos)
|
||||
_obj._addField(_fld)
|
||||
_pos = _fld.end
|
||||
_data = stream.read(1)
|
||||
return _obj
|
||||
fromStream = classmethod(fromStream)
|
||||
|
||||
## properties
|
||||
|
||||
year = property(lambda self: self.lastUpdate.year)
|
||||
month = property(lambda self: self.lastUpdate.month)
|
||||
day = property(lambda self: self.lastUpdate.day)
|
||||
|
||||
def ignoreErrors(self, value):
|
||||
"""Update `ignoreErrors` flag on self and all fields"""
|
||||
self._ignore_errors = value = bool(value)
|
||||
for _field in self.fields:
|
||||
_field.ignoreErrors = value
|
||||
ignoreErrors = property(
|
||||
lambda self: self._ignore_errors,
|
||||
ignoreErrors,
|
||||
doc="""Error processing mode for DBF field value conversion
|
||||
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
""")
|
||||
|
||||
## object representation
|
||||
|
||||
def __repr__(self):
|
||||
_rv = """\
|
||||
Version (signature): 0x%02x
|
||||
Last update: %s
|
||||
Header length: %d
|
||||
Record length: %d
|
||||
Record count: %d
|
||||
FieldName Type Len Dec
|
||||
""" % (self.signature, self.lastUpdate, self.headerLength,
|
||||
self.recordLength, self.recordCount)
|
||||
_rv += "\n".join(
|
||||
["%10s %4s %3s %3s" % _fld.fieldInfo() for _fld in self.fields]
|
||||
)
|
||||
return _rv
|
||||
|
||||
## internal methods
|
||||
|
||||
def _addField(self, *defs):
|
||||
"""Internal variant of the `addField` method.
|
||||
|
||||
This method doesn't set `self.changed` field to True.
|
||||
|
||||
Return value is a length of the appended records.
|
||||
Note: this method doesn't modify ``recordLength`` and
|
||||
``headerLength`` fields. Use `addField` instead of this
|
||||
method if you don't exactly know what you're doing.
|
||||
|
||||
"""
|
||||
# insure we have dbf.DbfFieldDef instances first (instantiation
|
||||
# from the tuple could raise an error, in such a case I don't
|
||||
# wanna add any of the definitions -- all will be ignored)
|
||||
_defs = []
|
||||
_recordLength = 0
|
||||
for _def in defs:
|
||||
if isinstance(_def, fields.DbfFieldDef):
|
||||
_obj = _def
|
||||
else:
|
||||
(_name, _type, _len, _dec) = (tuple(_def) + (None,) * 4)[:4]
|
||||
_cls = fields.lookupFor(_type)
|
||||
_obj = _cls(_name, _len, _dec,
|
||||
ignoreErrors=self._ignore_errors)
|
||||
_recordLength += _obj.length
|
||||
_defs.append(_obj)
|
||||
# and now extend field definitions and
|
||||
# update record length
|
||||
self.fields += _defs
|
||||
return _recordLength
|
||||
|
||||
## interface methods
|
||||
|
||||
def addField(self, *defs):
|
||||
"""Add field definition to the header.
|
||||
|
||||
Examples:
|
||||
dbfh.addField(
|
||||
("name", "C", 20),
|
||||
dbf.DbfCharacterFieldDef("surname", 20),
|
||||
dbf.DbfDateFieldDef("birthdate"),
|
||||
("member", "L"),
|
||||
)
|
||||
dbfh.addField(("price", "N", 5, 2))
|
||||
dbfh.addField(dbf.DbfNumericFieldDef("origprice", 5, 2))
|
||||
|
||||
"""
|
||||
_oldLen = self.recordLength
|
||||
self.recordLength += self._addField(*defs)
|
||||
if not _oldLen:
|
||||
self.recordLength += 1
|
||||
# XXX: may be just use:
|
||||
# self.recordeLength += self._addField(*defs) + bool(not _oldLen)
|
||||
# recalculate headerLength
|
||||
self.headerLength = 32 + (32 * len(self.fields)) + 1
|
||||
self.changed = True
|
||||
|
||||
def write(self, stream):
|
||||
"""Encode and write header to the stream."""
|
||||
stream.seek(0)
|
||||
stream.write(self.toString())
|
||||
stream.write("".join([_fld.toString() for _fld in self.fields]))
|
||||
stream.write(chr(0x0D)) # cr at end of all hdr data
|
||||
self.changed = False
|
||||
|
||||
def toString(self):
|
||||
"""Returned 32 chars length string with encoded header."""
|
||||
return struct.pack("<4BI2H",
|
||||
self.signature,
|
||||
self.year - 1900,
|
||||
self.month,
|
||||
self.day,
|
||||
self.recordCount,
|
||||
self.headerLength,
|
||||
self.recordLength) + "\0" * 20
|
||||
|
||||
def setCurrentDate(self):
|
||||
"""Update ``self.lastUpdate`` field with current date value."""
|
||||
self.lastUpdate = datetime.date.today()
|
||||
|
||||
def __getitem__(self, item):
|
||||
"""Return a field definition by numeric index or name string"""
|
||||
if isinstance(item, basestring):
|
||||
_name = item.upper()
|
||||
for _field in self.fields:
|
||||
if _field.name == _name:
|
||||
return _field
|
||||
else:
|
||||
raise KeyError(item)
|
||||
else:
|
||||
# item must be field index
|
||||
return self.fields[item]
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,262 @@
|
||||
"""DBF record definition.
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] __repr__: added special case for invalid field values
|
||||
10-feb-2007 [als] added .rawFromStream()
|
||||
30-oct-2006 [als] fix record length in .fromStream()
|
||||
04-jul-2006 [als] added export declaration
|
||||
20-dec-2005 [yc] DbfRecord.write() -> DbfRecord._write();
|
||||
added delete() method.
|
||||
16-dec-2005 [yc] record definition moved from `dbf`.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.7 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 09:05:49 $"[7:-2]
|
||||
|
||||
__all__ = ["DbfRecord"]
|
||||
|
||||
from itertools import izip
|
||||
|
||||
import utils
|
||||
|
||||
class DbfRecord(object):
|
||||
"""DBF record.
|
||||
|
||||
Instances of this class shouldn't be created manualy,
|
||||
use `dbf.Dbf.newRecord` instead.
|
||||
|
||||
Class implements mapping/sequence interface, so
|
||||
fields could be accessed via their names or indexes
|
||||
(names is a preffered way to access fields).
|
||||
|
||||
Hint:
|
||||
Use `store` method to save modified record.
|
||||
|
||||
Examples:
|
||||
Add new record to the database:
|
||||
db = Dbf(filename)
|
||||
rec = db.newRecord()
|
||||
rec["FIELD1"] = value1
|
||||
rec["FIELD2"] = value2
|
||||
rec.store()
|
||||
Or the same, but modify existed
|
||||
(second in this case) record:
|
||||
db = Dbf(filename)
|
||||
rec = db[2]
|
||||
rec["FIELD1"] = value1
|
||||
rec["FIELD2"] = value2
|
||||
rec.store()
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = "dbf", "index", "deleted", "fieldData"
|
||||
|
||||
## creation and initialization
|
||||
|
||||
def __init__(self, dbf, index=None, deleted=False, data=None):
|
||||
"""Instance initialiation.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance this record belonogs to.
|
||||
index:
|
||||
An integer record index or None. If this value is
|
||||
None, record will be appended to the DBF.
|
||||
deleted:
|
||||
Boolean flag indicating whether this record
|
||||
is a deleted record.
|
||||
data:
|
||||
A sequence or None. This is a data of the fields.
|
||||
If this argument is None, default values will be used.
|
||||
|
||||
"""
|
||||
self.dbf = dbf
|
||||
# XXX: I'm not sure ``index`` is necessary
|
||||
self.index = index
|
||||
self.deleted = deleted
|
||||
if data is None:
|
||||
self.fieldData = [_fd.defaultValue for _fd in dbf.header.fields]
|
||||
else:
|
||||
self.fieldData = list(data)
|
||||
|
||||
# XXX: validate self.index before calculating position?
|
||||
position = property(lambda self: self.dbf.header.headerLength + \
|
||||
self.index * self.dbf.header.recordLength)
|
||||
|
||||
def rawFromStream(cls, dbf, index):
|
||||
"""Return raw record contents read from the stream.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance containing the record.
|
||||
index:
|
||||
Index of the record in the records' container.
|
||||
This argument can't be None in this call.
|
||||
|
||||
Return value is a string containing record data in DBF format.
|
||||
|
||||
"""
|
||||
# XXX: may be write smth assuming, that current stream
|
||||
# position is the required one? it could save some
|
||||
# time required to calculate where to seek in the file
|
||||
dbf.stream.seek(dbf.header.headerLength +
|
||||
index * dbf.header.recordLength)
|
||||
return dbf.stream.read(dbf.header.recordLength)
|
||||
rawFromStream = classmethod(rawFromStream)
|
||||
|
||||
def fromStream(cls, dbf, index):
|
||||
"""Return a record read from the stream.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance new record should belong to.
|
||||
index:
|
||||
Index of the record in the records' container.
|
||||
This argument can't be None in this call.
|
||||
|
||||
Return value is an instance of the current class.
|
||||
|
||||
"""
|
||||
return cls.fromString(dbf, cls.rawFromStream(dbf, index), index)
|
||||
fromStream = classmethod(fromStream)
|
||||
|
||||
def fromString(cls, dbf, string, index=None):
|
||||
"""Return record read from the string object.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance new record should belong to.
|
||||
string:
|
||||
A string new record should be created from.
|
||||
index:
|
||||
Index of the record in the container. If this
|
||||
argument is None, record will be appended.
|
||||
|
||||
Return value is an instance of the current class.
|
||||
|
||||
"""
|
||||
return cls(dbf, index, string[0]=="*",
|
||||
[_fd.decodeFromRecord(string) for _fd in dbf.header.fields])
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
## object representation
|
||||
|
||||
def __repr__(self):
|
||||
_template = "%%%ds: %%s (%%s)" % max([len(_fld)
|
||||
for _fld in self.dbf.fieldNames])
|
||||
_rv = []
|
||||
for _fld in self.dbf.fieldNames:
|
||||
_val = self[_fld]
|
||||
if _val is utils.INVALID_VALUE:
|
||||
_rv.append(_template %
|
||||
(_fld, "None", "value cannot be decoded"))
|
||||
else:
|
||||
_rv.append(_template % (_fld, _val, type(_val)))
|
||||
return "\n".join(_rv)
|
||||
|
||||
## protected methods
|
||||
|
||||
def _write(self):
|
||||
"""Write data to the dbf stream.
|
||||
|
||||
Note:
|
||||
This isn't a public method, it's better to
|
||||
use 'store' instead publically.
|
||||
Be design ``_write`` method should be called
|
||||
only from the `Dbf` instance.
|
||||
|
||||
|
||||
"""
|
||||
self._validateIndex(False)
|
||||
self.dbf.stream.seek(self.position)
|
||||
self.dbf.stream.write(self.toString())
|
||||
# FIXME: may be move this write somewhere else?
|
||||
# why we should check this condition for each record?
|
||||
if self.index == len(self.dbf):
|
||||
# this is the last record,
|
||||
# we should write SUB (ASCII 26)
|
||||
self.dbf.stream.write("\x1A")
|
||||
|
||||
## utility methods
|
||||
|
||||
def _validateIndex(self, allowUndefined=True, checkRange=False):
|
||||
"""Valid ``self.index`` value.
|
||||
|
||||
If ``allowUndefined`` argument is True functions does nothing
|
||||
in case of ``self.index`` pointing to None object.
|
||||
|
||||
"""
|
||||
if self.index is None:
|
||||
if not allowUndefined:
|
||||
raise ValueError("Index is undefined")
|
||||
elif self.index < 0:
|
||||
raise ValueError("Index can't be negative (%s)" % self.index)
|
||||
elif checkRange and self.index <= self.dbf.header.recordCount:
|
||||
raise ValueError("There are only %d records in the DBF" %
|
||||
self.dbf.header.recordCount)
|
||||
|
||||
## interface methods
|
||||
|
||||
def store(self):
|
||||
"""Store current record in the DBF.
|
||||
|
||||
If ``self.index`` is None, this record will be appended to the
|
||||
records of the DBF this records belongs to; or replaced otherwise.
|
||||
|
||||
"""
|
||||
self._validateIndex()
|
||||
if self.index is None:
|
||||
self.index = len(self.dbf)
|
||||
self.dbf.append(self)
|
||||
else:
|
||||
self.dbf[self.index] = self
|
||||
|
||||
def delete(self):
|
||||
"""Mark method as deleted."""
|
||||
self.deleted = True
|
||||
|
||||
def toString(self):
|
||||
"""Return string packed record values."""
|
||||
return "".join([" *"[self.deleted]] + [
|
||||
_def.encodeValue(_dat)
|
||||
for (_def, _dat) in izip(self.dbf.header.fields, self.fieldData)
|
||||
])
|
||||
|
||||
def asList(self):
|
||||
"""Return a flat list of fields.
|
||||
|
||||
Note:
|
||||
Change of the list's values won't change
|
||||
real values stored in this object.
|
||||
|
||||
"""
|
||||
return self.fieldData[:]
|
||||
|
||||
def asDict(self):
|
||||
"""Return a dictionary of fields.
|
||||
|
||||
Note:
|
||||
Change of the dicts's values won't change
|
||||
real values stored in this object.
|
||||
|
||||
"""
|
||||
return dict([_i for _i in izip(self.dbf.fieldNames, self.fieldData)])
|
||||
|
||||
def __getitem__(self, key):
|
||||
"""Return value by field name or field index."""
|
||||
if isinstance(key, (long, int)):
|
||||
# integer index of the field
|
||||
return self.fieldData[key]
|
||||
# assuming string field name
|
||||
return self.fieldData[self.dbf.indexOfFieldName(key)]
|
||||
|
||||
def __setitem__(self, key, value):
|
||||
"""Set field value by integer index of the field or string name."""
|
||||
if isinstance(key, (int, long)):
|
||||
# integer index of the field
|
||||
return self.fieldData[key]
|
||||
# assuming string field name
|
||||
self.fieldData[self.dbf.indexOfFieldName(key)] = value
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,170 @@
|
||||
"""String utilities.
|
||||
|
||||
TODO:
|
||||
- allow strings in getDateTime routine;
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] added INVALID_VALUE
|
||||
10-feb-2007 [als] allow date strings padded with spaces instead of zeroes
|
||||
20-dec-2005 [yc] handle long objects in getDate/getDateTime
|
||||
16-dec-2005 [yc] created from ``strutil`` module.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.4 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 08:57:17 $"[7:-2]
|
||||
|
||||
import datetime
|
||||
import time
|
||||
|
||||
|
||||
def unzfill(str):
|
||||
"""Return a string without ASCII NULs.
|
||||
|
||||
This function searchers for the first NUL (ASCII 0) occurance
|
||||
and truncates string till that position.
|
||||
|
||||
"""
|
||||
try:
|
||||
return str[:str.index('\0')]
|
||||
except ValueError:
|
||||
return str
|
||||
|
||||
|
||||
def getDate(date=None):
|
||||
"""Return `datetime.date` instance.
|
||||
|
||||
Type of the ``date`` argument could be one of the following:
|
||||
None:
|
||||
use current date value;
|
||||
datetime.date:
|
||||
this value will be returned;
|
||||
datetime.datetime:
|
||||
the result of the date.date() will be returned;
|
||||
string:
|
||||
assuming "%Y%m%d" or "%y%m%dd" format;
|
||||
number:
|
||||
assuming it's a timestamp (returned for example
|
||||
by the time.time() call;
|
||||
sequence:
|
||||
assuming (year, month, day, ...) sequence;
|
||||
|
||||
Additionaly, if ``date`` has callable ``ticks`` attribute,
|
||||
it will be used and result of the called would be treated
|
||||
as a timestamp value.
|
||||
|
||||
"""
|
||||
if date is None:
|
||||
# use current value
|
||||
return datetime.date.today()
|
||||
if isinstance(date, datetime.date):
|
||||
return date
|
||||
if isinstance(date, datetime.datetime):
|
||||
return date.date()
|
||||
if isinstance(date, (int, long, float)):
|
||||
# date is a timestamp
|
||||
return datetime.date.fromtimestamp(date)
|
||||
if isinstance(date, basestring):
|
||||
date = date.replace(" ", "0")
|
||||
if len(date) == 6:
|
||||
# yymmdd
|
||||
return datetime.date(*time.strptime(date, "%y%m%d")[:3])
|
||||
# yyyymmdd
|
||||
return datetime.date(*time.strptime(date, "%Y%m%d")[:3])
|
||||
if hasattr(date, "__getitem__"):
|
||||
# a sequence (assuming date/time tuple)
|
||||
return datetime.date(*date[:3])
|
||||
return datetime.date.fromtimestamp(date.ticks())
|
||||
|
||||
|
||||
def getDateTime(value=None):
|
||||
"""Return `datetime.datetime` instance.
|
||||
|
||||
Type of the ``value`` argument could be one of the following:
|
||||
None:
|
||||
use current date value;
|
||||
datetime.date:
|
||||
result will be converted to the `datetime.datetime` instance
|
||||
using midnight;
|
||||
datetime.datetime:
|
||||
``value`` will be returned as is;
|
||||
string:
|
||||
*** CURRENTLY NOT SUPPORTED ***;
|
||||
number:
|
||||
assuming it's a timestamp (returned for example
|
||||
by the time.time() call;
|
||||
sequence:
|
||||
assuming (year, month, day, ...) sequence;
|
||||
|
||||
Additionaly, if ``value`` has callable ``ticks`` attribute,
|
||||
it will be used and result of the called would be treated
|
||||
as a timestamp value.
|
||||
|
||||
"""
|
||||
if value is None:
|
||||
# use current value
|
||||
return datetime.datetime.today()
|
||||
if isinstance(value, datetime.datetime):
|
||||
return value
|
||||
if isinstance(value, datetime.date):
|
||||
return datetime.datetime.fromordinal(value.toordinal())
|
||||
if isinstance(value, (int, long, float)):
|
||||
# value is a timestamp
|
||||
return datetime.datetime.fromtimestamp(value)
|
||||
if isinstance(value, basestring):
|
||||
raise NotImplementedError("Strings aren't currently implemented")
|
||||
if hasattr(value, "__getitem__"):
|
||||
# a sequence (assuming date/time tuple)
|
||||
return datetime.datetime(*tuple(value)[:6])
|
||||
return datetime.datetime.fromtimestamp(value.ticks())
|
||||
|
||||
|
||||
class classproperty(property):
|
||||
"""Works in the same way as a ``property``, but for the classes."""
|
||||
|
||||
def __get__(self, obj, cls):
|
||||
return self.fget(cls)
|
||||
|
||||
|
||||
class _InvalidValue(object):
|
||||
|
||||
"""Value returned from DBF records when field validation fails
|
||||
|
||||
The value is not equal to anything except for itself
|
||||
and equal to all empty values: None, 0, empty string etc.
|
||||
In other words, invalid value is equal to None and not equal
|
||||
to None at the same time.
|
||||
|
||||
This value yields zero upon explicit conversion to a number type,
|
||||
empty string for string types, and False for boolean.
|
||||
|
||||
"""
|
||||
|
||||
def __eq__(self, other):
|
||||
return not other
|
||||
|
||||
def __ne__(self, other):
|
||||
return not (other is self)
|
||||
|
||||
def __nonzero__(self):
|
||||
return False
|
||||
|
||||
def __int__(self):
|
||||
return 0
|
||||
__long__ = __int__
|
||||
|
||||
def __float__(self):
|
||||
return 0.0
|
||||
|
||||
def __str__(self):
|
||||
return ""
|
||||
|
||||
def __unicode__(self):
|
||||
return u""
|
||||
|
||||
def __repr__(self):
|
||||
return "<INVALID>"
|
||||
|
||||
# invalid value is a constant singleton
|
||||
INVALID_VALUE = _InvalidValue()
|
||||
|
||||
# vim: set et sts=4 sw=4 :
|
||||
@@ -0,0 +1,293 @@
|
||||
#! /usr/bin/env python
|
||||
"""DBF accessing helpers.
|
||||
|
||||
FIXME: more documentation needed
|
||||
|
||||
Examples:
|
||||
|
||||
Create new table, setup structure, add records:
|
||||
|
||||
dbf = Dbf(filename, new=True)
|
||||
dbf.addField(
|
||||
("NAME", "C", 15),
|
||||
("SURNAME", "C", 25),
|
||||
("INITIALS", "C", 10),
|
||||
("BIRTHDATE", "D"),
|
||||
)
|
||||
for (n, s, i, b) in (
|
||||
("John", "Miller", "YC", (1980, 10, 11)),
|
||||
("Andy", "Larkin", "", (1980, 4, 11)),
|
||||
):
|
||||
rec = dbf.newRecord()
|
||||
rec["NAME"] = n
|
||||
rec["SURNAME"] = s
|
||||
rec["INITIALS"] = i
|
||||
rec["BIRTHDATE"] = b
|
||||
rec.store()
|
||||
dbf.close()
|
||||
|
||||
Open existed dbf, read some data:
|
||||
|
||||
dbf = Dbf(filename, True)
|
||||
for rec in dbf:
|
||||
for fldName in dbf.fieldNames:
|
||||
print '%s:\t %s (%s)' % (fldName, rec[fldName],
|
||||
type(rec[fldName]))
|
||||
print
|
||||
dbf.close()
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] export INVALID_VALUE;
|
||||
Dbf: added .ignoreErrors, .INVALID_VALUE
|
||||
04-jul-2006 [als] added export declaration
|
||||
20-dec-2005 [yc] removed fromStream and newDbf methods:
|
||||
use argument of __init__ call must be used instead;
|
||||
added class fields pointing to the header and
|
||||
record classes.
|
||||
17-dec-2005 [yc] split to several modules; reimplemented
|
||||
13-dec-2005 [yc] adapted to the changes of the `strutil` module.
|
||||
13-sep-2002 [als] support FoxPro Timestamp datatype
|
||||
15-nov-1999 [jjk] documentation updates, add demo
|
||||
24-aug-1998 [jjk] add some encodeValue methods (not tested), other tweaks
|
||||
08-jun-1998 [jjk] fix problems, add more features
|
||||
20-feb-1998 [jjk] fix problems, add more features
|
||||
19-feb-1998 [jjk] add create/write capabilities
|
||||
18-feb-1998 [jjk] from dbfload.py
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.7 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 09:23:13 $"[7:-2]
|
||||
__author__ = "Jeff Kunce <kuncej@mail.conservation.state.mo.us>"
|
||||
|
||||
__all__ = ["Dbf"]
|
||||
|
||||
from . import header
|
||||
from . import record
|
||||
from .utils import INVALID_VALUE
|
||||
|
||||
class Dbf(object):
|
||||
"""DBF accessor.
|
||||
|
||||
FIXME:
|
||||
docs and examples needed (dont' forget to tell
|
||||
about problems adding new fields on the fly)
|
||||
|
||||
Implementation notes:
|
||||
``_new`` field is used to indicate whether this is
|
||||
a new data table. `addField` could be used only for
|
||||
the new tables! If at least one record was appended
|
||||
to the table it's structure couldn't be changed.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("name", "header", "stream",
|
||||
"_changed", "_new", "_ignore_errors")
|
||||
|
||||
HeaderClass = header.DbfHeader
|
||||
RecordClass = record.DbfRecord
|
||||
INVALID_VALUE = INVALID_VALUE
|
||||
|
||||
## initialization and creation helpers
|
||||
|
||||
def __init__(self, f, readOnly=False, new=False, ignoreErrors=False):
|
||||
"""Initialize instance.
|
||||
|
||||
Arguments:
|
||||
f:
|
||||
Filename or file-like object.
|
||||
new:
|
||||
True if new data table must be created. Assume
|
||||
data table exists if this argument is False.
|
||||
readOnly:
|
||||
if ``f`` argument is a string file will
|
||||
be opend in read-only mode; in other cases
|
||||
this argument is ignored. This argument is ignored
|
||||
even if ``new`` argument is True.
|
||||
headerObj:
|
||||
`header.DbfHeader` instance or None. If this argument
|
||||
is None, new empty header will be used with the
|
||||
all fields set by default.
|
||||
ignoreErrors:
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
"""
|
||||
if isinstance(f, str):
|
||||
# a filename
|
||||
self.name = f
|
||||
if new:
|
||||
# new table (table file must be
|
||||
# created or opened and truncated)
|
||||
self.stream = open(f, "w+b")
|
||||
else:
|
||||
# tabe file must exist
|
||||
self.stream = open(f, ("r+b", "rb")[bool(readOnly)])
|
||||
else:
|
||||
# a stream
|
||||
self.name = getattr(f, "name", "")
|
||||
self.stream = f
|
||||
if new:
|
||||
# if this is a new table, header will be empty
|
||||
self.header = self.HeaderClass()
|
||||
else:
|
||||
# or instantiated using stream
|
||||
self.header = self.HeaderClass.fromStream(self.stream)
|
||||
self.ignoreErrors = ignoreErrors
|
||||
self._new = bool(new)
|
||||
self._changed = False
|
||||
|
||||
## properties
|
||||
|
||||
closed = property(lambda self: self.stream.closed)
|
||||
recordCount = property(lambda self: self.header.recordCount)
|
||||
fieldNames = property(
|
||||
lambda self: [_fld.name for _fld in self.header.fields])
|
||||
fieldDefs = property(lambda self: self.header.fields)
|
||||
changed = property(lambda self: self._changed or self.header.changed)
|
||||
|
||||
def ignoreErrors(self, value):
|
||||
"""Update `ignoreErrors` flag on the header object and self"""
|
||||
self.header.ignoreErrors = self._ignore_errors = bool(value)
|
||||
ignoreErrors = property(
|
||||
lambda self: self._ignore_errors,
|
||||
ignoreErrors,
|
||||
doc="""Error processing mode for DBF field value conversion
|
||||
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
""")
|
||||
|
||||
## protected methods
|
||||
|
||||
def _fixIndex(self, index):
|
||||
"""Return fixed index.
|
||||
|
||||
This method fails if index isn't a numeric object
|
||||
(long or int). Or index isn't in a valid range
|
||||
(less or equal to the number of records in the db).
|
||||
|
||||
If ``index`` is a negative number, it will be
|
||||
treated as a negative indexes for list objects.
|
||||
|
||||
Return:
|
||||
Return value is numeric object maning valid index.
|
||||
|
||||
"""
|
||||
if not isinstance(index, int):
|
||||
raise TypeError("Index must be a numeric object")
|
||||
if index < 0:
|
||||
# index from the right side
|
||||
# fix it to the left-side index
|
||||
index += len(self) + 1
|
||||
if index >= len(self):
|
||||
raise IndexError("Record index out of range")
|
||||
return index
|
||||
|
||||
## iterface methods
|
||||
|
||||
def close(self):
|
||||
self.flush()
|
||||
self.stream.close()
|
||||
|
||||
def flush(self):
|
||||
"""Flush data to the associated stream."""
|
||||
if self.changed:
|
||||
self.header.setCurrentDate()
|
||||
self.header.write(self.stream)
|
||||
self.stream.flush()
|
||||
self._changed = False
|
||||
|
||||
def indexOfFieldName(self, name):
|
||||
"""Index of field named ``name``."""
|
||||
# FIXME: move this to header class
|
||||
names = [f.name for f in self.header.fields]
|
||||
return names.index(name.upper())
|
||||
|
||||
def newRecord(self):
|
||||
"""Return new record, which belong to this table."""
|
||||
return self.RecordClass(self)
|
||||
|
||||
def append(self, record):
|
||||
"""Append ``record`` to the database."""
|
||||
record.index = self.header.recordCount
|
||||
record._write()
|
||||
self.header.recordCount += 1
|
||||
self._changed = True
|
||||
self._new = False
|
||||
|
||||
def addField(self, *defs):
|
||||
"""Add field definitions.
|
||||
|
||||
For more information see `header.DbfHeader.addField`.
|
||||
|
||||
"""
|
||||
if self._new:
|
||||
self.header.addField(*defs)
|
||||
else:
|
||||
raise TypeError("At least one record was added, "
|
||||
"structure can't be changed")
|
||||
|
||||
## 'magic' methods (representation and sequence interface)
|
||||
|
||||
def __repr__(self):
|
||||
return "Dbf stream '%s'\n" % self.stream + repr(self.header)
|
||||
|
||||
def __len__(self):
|
||||
"""Return number of records."""
|
||||
return self.recordCount
|
||||
|
||||
def __getitem__(self, index):
|
||||
"""Return `DbfRecord` instance."""
|
||||
return self.RecordClass.fromStream(self, self._fixIndex(index))
|
||||
|
||||
def __setitem__(self, index, record):
|
||||
"""Write `DbfRecord` instance to the stream."""
|
||||
record.index = self._fixIndex(index)
|
||||
record._write()
|
||||
self._changed = True
|
||||
self._new = False
|
||||
|
||||
#def __del__(self):
|
||||
# """Flush stream upon deletion of the object."""
|
||||
# self.flush()
|
||||
|
||||
|
||||
def demoRead(filename):
|
||||
_dbf = Dbf(filename, True)
|
||||
for _rec in _dbf:
|
||||
print()
|
||||
print(repr(_rec))
|
||||
_dbf.close()
|
||||
|
||||
def demoCreate(filename):
|
||||
_dbf = Dbf(filename, new=True)
|
||||
_dbf.addField(
|
||||
("NAME", "C", 15),
|
||||
("SURNAME", "C", 25),
|
||||
("INITIALS", "C", 10),
|
||||
("BIRTHDATE", "D"),
|
||||
)
|
||||
for (_n, _s, _i, _b) in (
|
||||
("John", "Miller", "YC", (1981, 1, 2)),
|
||||
("Andy", "Larkin", "AL", (1982, 3, 4)),
|
||||
("Bill", "Clinth", "", (1983, 5, 6)),
|
||||
("Bobb", "McNail", "", (1984, 7, 8)),
|
||||
):
|
||||
_rec = _dbf.newRecord()
|
||||
_rec["NAME"] = _n
|
||||
_rec["SURNAME"] = _s
|
||||
_rec["INITIALS"] = _i
|
||||
_rec["BIRTHDATE"] = _b
|
||||
_rec.store()
|
||||
print(repr(_dbf))
|
||||
_dbf.close()
|
||||
|
||||
if (__name__=='__main__'):
|
||||
import sys
|
||||
_name = len(sys.argv) > 1 and sys.argv[1] or "county.dbf"
|
||||
demoCreate(_name)
|
||||
demoRead(_name)
|
||||
|
||||
# vim: set et sw=4 sts=4 :
|
||||
@@ -0,0 +1,182 @@
|
||||
#!/usr/bin/python
|
||||
""".DBF creation helpers.
|
||||
|
||||
Note: this is a legacy interface. New code should use Dbf class
|
||||
for table creation (see examples in dbf.py)
|
||||
|
||||
TODO:
|
||||
- handle Memo fields.
|
||||
- check length of the fields accoring to the
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
|
||||
"""
|
||||
"""History (most recent first)
|
||||
04-jul-2006 [als] added export declaration;
|
||||
updated for dbfpy 2.0
|
||||
15-dec-2005 [yc] define dbf_new.__slots__
|
||||
14-dec-2005 [yc] added vim modeline; retab'd; added doc-strings;
|
||||
dbf_new now is a new class (inherited from object)
|
||||
??-jun-2000 [--] added by Hans Fiby
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.4 $"[11:-2]
|
||||
__date__ = "$Date: 2006/07/04 08:18:18 $"[7:-2]
|
||||
|
||||
__all__ = ["dbf_new"]
|
||||
|
||||
from .dbf import *
|
||||
from .fields import *
|
||||
from .header import *
|
||||
from .record import *
|
||||
|
||||
class _FieldDefinition(object):
|
||||
"""Field definition.
|
||||
|
||||
This is a simple structure, which contains ``name``, ``type``,
|
||||
``len``, ``dec`` and ``cls`` fields.
|
||||
|
||||
Objects also implement get/setitem magic functions, so fields
|
||||
could be accessed via sequence iterface, where 'name' has
|
||||
index 0, 'type' index 1, 'len' index 2, 'dec' index 3 and
|
||||
'cls' could be located at index 4.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = "name", "type", "len", "dec", "cls"
|
||||
|
||||
# WARNING: be attentive - dictionaries are mutable!
|
||||
FLD_TYPES = {
|
||||
# type: (cls, len)
|
||||
"C": (DbfCharacterFieldDef, None),
|
||||
"N": (DbfNumericFieldDef, None),
|
||||
"L": (DbfLogicalFieldDef, 1),
|
||||
# FIXME: support memos
|
||||
# "M": (DbfMemoFieldDef),
|
||||
"D": (DbfDateFieldDef, 8),
|
||||
# FIXME: I'm not sure length should be 14 characters!
|
||||
# but temporary I use it, cuz date is 8 characters
|
||||
# and time 6 (hhmmss)
|
||||
"T": (DbfDateTimeFieldDef, 14),
|
||||
}
|
||||
|
||||
def __init__(self, name, type, len=None, dec=0):
|
||||
_cls, _len = self.FLD_TYPES[type]
|
||||
if _len is None:
|
||||
if len is None:
|
||||
raise ValueError("Field length must be defined")
|
||||
_len = len
|
||||
self.name = name
|
||||
self.type = type
|
||||
self.len = _len
|
||||
self.dec = dec
|
||||
self.cls = _cls
|
||||
|
||||
def getDbfField(self):
|
||||
"Return `DbfFieldDef` instance from the current definition."
|
||||
return self.cls(self.name, self.len, self.dec)
|
||||
|
||||
def appendToHeader(self, dbfh):
|
||||
"""Create a `DbfFieldDef` instance and append it to the dbf header.
|
||||
|
||||
Arguments:
|
||||
dbfh: `DbfHeader` instance.
|
||||
|
||||
"""
|
||||
_dbff = self.getDbfField()
|
||||
dbfh.addField(_dbff)
|
||||
|
||||
|
||||
class dbf_new(object):
|
||||
"""New .DBF creation helper.
|
||||
|
||||
Example Usage:
|
||||
|
||||
dbfn = dbf_new()
|
||||
dbfn.add_field("name",'C',80)
|
||||
dbfn.add_field("price",'N',10,2)
|
||||
dbfn.add_field("date",'D',8)
|
||||
dbfn.write("tst.dbf")
|
||||
|
||||
Note:
|
||||
This module cannot handle Memo-fields,
|
||||
they are special.
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("fields",)
|
||||
|
||||
FieldDefinitionClass = _FieldDefinition
|
||||
|
||||
def __init__(self):
|
||||
self.fields = []
|
||||
|
||||
def add_field(self, name, typ, len, dec=0):
|
||||
"""Add field definition.
|
||||
|
||||
Arguments:
|
||||
name:
|
||||
field name (str object). field name must not
|
||||
contain ASCII NULs and it's length shouldn't
|
||||
exceed 10 characters.
|
||||
typ:
|
||||
type of the field. this must be a single character
|
||||
from the "CNLMDT" set meaning character, numeric,
|
||||
logical, memo, date and date/time respectively.
|
||||
len:
|
||||
length of the field. this argument is used only for
|
||||
the character and numeric fields. all other fields
|
||||
have fixed length.
|
||||
FIXME: use None as a default for this argument?
|
||||
dec:
|
||||
decimal precision. used only for the numric fields.
|
||||
|
||||
"""
|
||||
self.fields.append(self.FieldDefinitionClass(name, typ, len, dec))
|
||||
|
||||
def write(self, filename):
|
||||
"""Create empty .DBF file using current structure."""
|
||||
_dbfh = DbfHeader()
|
||||
_dbfh.setCurrentDate()
|
||||
for _fldDef in self.fields:
|
||||
_fldDef.appendToHeader(_dbfh)
|
||||
|
||||
_dbfStream = open(filename, "wb")
|
||||
_dbfh.write(_dbfStream)
|
||||
_dbfStream.close()
|
||||
|
||||
|
||||
if (__name__=='__main__'):
|
||||
# create a new DBF-File
|
||||
dbfn=dbf_new()
|
||||
dbfn.add_field("name",'C',80)
|
||||
dbfn.add_field("price",'N',10,2)
|
||||
dbfn.add_field("date",'D',8)
|
||||
dbfn.write("tst.dbf")
|
||||
# test new dbf
|
||||
print("*** created tst.dbf: ***")
|
||||
dbft = Dbf('tst.dbf', readOnly=0)
|
||||
print(repr(dbft))
|
||||
# add a record
|
||||
rec=DbfRecord(dbft)
|
||||
rec['name']='something'
|
||||
rec['price']=10.5
|
||||
rec['date']=(2000,1,12)
|
||||
rec.store()
|
||||
# add another record
|
||||
rec=DbfRecord(dbft)
|
||||
rec['name']='foo and bar'
|
||||
rec['price']=12234
|
||||
rec['date']=(1992,7,15)
|
||||
rec.store()
|
||||
|
||||
# show the records
|
||||
print("*** inserted 2 records into tst.dbf: ***")
|
||||
print(repr(dbft))
|
||||
for i1 in range(len(dbft)):
|
||||
rec = dbft[i1]
|
||||
for fldName in dbft.fieldNames:
|
||||
print('%s:\t %s'%(fldName, rec[fldName]))
|
||||
print()
|
||||
dbft.close()
|
||||
|
||||
# vim: set et sts=4 sw=4 :
|
||||
@@ -0,0 +1,467 @@
|
||||
"""DBF fields definitions.
|
||||
|
||||
TODO:
|
||||
- make memos work
|
||||
"""
|
||||
"""History (most recent first):
|
||||
26-may-2009 [als] DbfNumericFieldDef.decodeValue: strip zero bytes
|
||||
05-feb-2009 [als] DbfDateFieldDef.encodeValue: empty arg produces empty date
|
||||
16-sep-2008 [als] DbfNumericFieldDef decoding looks for decimal point
|
||||
in the value to select float or integer return type
|
||||
13-mar-2008 [als] check field name length in constructor
|
||||
11-feb-2007 [als] handle value conversion errors
|
||||
10-feb-2007 [als] DbfFieldDef: added .rawFromRecord()
|
||||
01-dec-2006 [als] Timestamp columns use None for empty values
|
||||
31-oct-2006 [als] support field types 'F' (float), 'I' (integer)
|
||||
and 'Y' (currency);
|
||||
automate export and registration of field classes
|
||||
04-jul-2006 [als] added export declaration
|
||||
10-mar-2006 [als] decode empty values for Date and Logical fields;
|
||||
show field name in errors
|
||||
10-mar-2006 [als] fix Numeric value decoding: according to spec,
|
||||
value always is string representation of the number;
|
||||
ensure that encoded Numeric value fits into the field
|
||||
20-dec-2005 [yc] use field names in upper case
|
||||
15-dec-2005 [yc] field definitions moved from `dbf`.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.14 $"[11:-2]
|
||||
__date__ = "$Date: 2009/05/26 05:16:51 $"[7:-2]
|
||||
|
||||
__all__ = ["lookupFor",] # field classes added at the end of the module
|
||||
|
||||
import datetime
|
||||
import struct
|
||||
import sys
|
||||
|
||||
from . import utils
|
||||
|
||||
## abstract definitions
|
||||
|
||||
class DbfFieldDef(object):
|
||||
"""Abstract field definition.
|
||||
|
||||
Child classes must override ``type`` class attribute to provide datatype
|
||||
infromation of the field definition. For more info about types visit
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
|
||||
Also child classes must override ``defaultValue`` field to provide
|
||||
default value for the field value.
|
||||
|
||||
If child class has fixed length ``length`` class attribute must be
|
||||
overriden and set to the valid value. None value means, that field
|
||||
isn't of fixed length.
|
||||
|
||||
Note: ``name`` field must not be changed after instantiation.
|
||||
|
||||
"""
|
||||
|
||||
|
||||
__slots__ = ("name", "decimalCount",
|
||||
"start", "end", "ignoreErrors")
|
||||
|
||||
# length of the field, None in case of variable-length field,
|
||||
# or a number if this field is a fixed-length field
|
||||
length = None
|
||||
|
||||
# field type. for more information about fields types visit
|
||||
# `http://www.clicketyclick.dk/databases/xbase/format/data_types.html`
|
||||
# must be overriden in child classes
|
||||
typeCode = None
|
||||
|
||||
# default value for the field. this field must be
|
||||
# overriden in child classes
|
||||
defaultValue = None
|
||||
|
||||
def __init__(self, name, length=None, decimalCount=None,
|
||||
start=None, stop=None, ignoreErrors=False,
|
||||
):
|
||||
"""Initialize instance."""
|
||||
assert self.typeCode is not None, "Type code must be overriden"
|
||||
assert self.defaultValue is not None, "Default value must be overriden"
|
||||
## fix arguments
|
||||
if len(name) >10:
|
||||
raise ValueError("Field name \"%s\" is too long" % name)
|
||||
name = str(name).upper()
|
||||
if self.__class__.length is None:
|
||||
if length is None:
|
||||
raise ValueError("[%s] Length isn't specified" % name)
|
||||
length = int(length)
|
||||
if length <= 0:
|
||||
raise ValueError("[%s] Length must be a positive integer"
|
||||
% name)
|
||||
else:
|
||||
length = self.length
|
||||
if decimalCount is None:
|
||||
decimalCount = 0
|
||||
## set fields
|
||||
self.name = name
|
||||
# FIXME: validate length according to the specification at
|
||||
# http://www.clicketyclick.dk/databases/xbase/format/data_types.html
|
||||
self.length = length
|
||||
self.decimalCount = decimalCount
|
||||
self.ignoreErrors = ignoreErrors
|
||||
self.start = start
|
||||
self.end = stop
|
||||
|
||||
def __cmp__(self, other):
|
||||
return cmp(self.name, str(other).upper())
|
||||
|
||||
def __hash__(self):
|
||||
return hash(self.name)
|
||||
|
||||
def fromString(cls, string, start, ignoreErrors=False):
|
||||
"""Decode dbf field definition from the string data.
|
||||
|
||||
Arguments:
|
||||
string:
|
||||
a string, dbf definition is decoded from. length of
|
||||
the string must be 32 bytes.
|
||||
start:
|
||||
position in the database file.
|
||||
ignoreErrors:
|
||||
initial error processing mode for the new field (boolean)
|
||||
|
||||
"""
|
||||
assert len(string) == 32
|
||||
_length = string[16]
|
||||
return cls(utils.unzfill(string)[:11].decode('utf-8'), _length,
|
||||
string[17], start, start + _length, ignoreErrors=ignoreErrors)
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
def toString(self):
|
||||
"""Return encoded field definition.
|
||||
|
||||
Return:
|
||||
Return value is a string object containing encoded
|
||||
definition of this field.
|
||||
|
||||
"""
|
||||
if sys.version_info < (2, 4):
|
||||
# earlier versions did not support padding character
|
||||
_name = self.name[:11] + "\0" * (11 - len(self.name))
|
||||
else:
|
||||
_name = self.name.ljust(11, '\0')
|
||||
return (
|
||||
_name +
|
||||
self.typeCode +
|
||||
#data address
|
||||
chr(0) * 4 +
|
||||
chr(self.length) +
|
||||
chr(self.decimalCount) +
|
||||
chr(0) * 14
|
||||
)
|
||||
|
||||
def __repr__(self):
|
||||
return "%-10s %1s %3d %3d" % self.fieldInfo()
|
||||
|
||||
def fieldInfo(self):
|
||||
"""Return field information.
|
||||
|
||||
Return:
|
||||
Return value is a (name, type, length, decimals) tuple.
|
||||
|
||||
"""
|
||||
return (self.name, self.typeCode, self.length, self.decimalCount)
|
||||
|
||||
def rawFromRecord(self, record):
|
||||
"""Return a "raw" field value from the record string."""
|
||||
return record[self.start:self.end]
|
||||
|
||||
def decodeFromRecord(self, record):
|
||||
"""Return decoded field value from the record string."""
|
||||
try:
|
||||
return self.decodeValue(self.rawFromRecord(record))
|
||||
except:
|
||||
if self.ignoreErrors:
|
||||
return utils.INVALID_VALUE
|
||||
else:
|
||||
raise
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return decoded value from string value.
|
||||
|
||||
This method shouldn't be used publicly. It's called from the
|
||||
`decodeFromRecord` method.
|
||||
|
||||
This is an abstract method and it must be overridden in child classes.
|
||||
"""
|
||||
raise NotImplementedError
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return str object containing encoded field value.
|
||||
|
||||
This is an abstract method and it must be overriden in child classes.
|
||||
"""
|
||||
raise NotImplementedError
|
||||
|
||||
## real classes
|
||||
|
||||
class DbfCharacterFieldDef(DbfFieldDef):
|
||||
"""Definition of the character field."""
|
||||
|
||||
typeCode = "C"
|
||||
defaultValue = b''
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return string object.
|
||||
|
||||
Return value is a ``value`` argument with stripped right spaces.
|
||||
|
||||
"""
|
||||
return value.rstrip(b' ').decode('utf-8')
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return raw data string encoded from a ``value``."""
|
||||
return str(value)[:self.length].ljust(self.length)
|
||||
|
||||
|
||||
class DbfNumericFieldDef(DbfFieldDef):
|
||||
"""Definition of the numeric field."""
|
||||
|
||||
typeCode = "N"
|
||||
# XXX: now I'm not sure it was a good idea to make a class field
|
||||
# `defaultValue` instead of a generic method as it was implemented
|
||||
# previously -- it's ok with all types except number, cuz
|
||||
# if self.decimalCount is 0, we should return 0 and 0.0 otherwise.
|
||||
defaultValue = 0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a number decoded from ``value``.
|
||||
|
||||
If decimals is zero, value will be decoded as an integer;
|
||||
or as a float otherwise.
|
||||
|
||||
Return:
|
||||
Return value is a int (long) or float instance.
|
||||
|
||||
"""
|
||||
value = value.strip(b' \0')
|
||||
if b'.' in value:
|
||||
# a float (has decimal separator)
|
||||
return float(value)
|
||||
elif value:
|
||||
# must be an integer
|
||||
return int(value)
|
||||
else:
|
||||
return 0
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
_rv = ("%*.*f" % (self.length, self.decimalCount, value))
|
||||
if len(_rv) > self.length:
|
||||
_ppos = _rv.find(".")
|
||||
if 0 <= _ppos <= self.length:
|
||||
_rv = _rv[:self.length]
|
||||
else:
|
||||
raise ValueError("[%s] Numeric overflow: %s (field width: %i)"
|
||||
% (self.name, _rv, self.length))
|
||||
return _rv
|
||||
|
||||
class DbfFloatFieldDef(DbfNumericFieldDef):
|
||||
"""Definition of the float field - same as numeric."""
|
||||
|
||||
typeCode = "F"
|
||||
|
||||
class DbfIntegerFieldDef(DbfFieldDef):
|
||||
"""Definition of the integer field."""
|
||||
|
||||
typeCode = "I"
|
||||
length = 4
|
||||
defaultValue = 0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return an integer number decoded from ``value``."""
|
||||
return struct.unpack("<i", value)[0]
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
return struct.pack("<i", int(value))
|
||||
|
||||
class DbfCurrencyFieldDef(DbfFieldDef):
|
||||
"""Definition of the currency field."""
|
||||
|
||||
typeCode = "Y"
|
||||
length = 8
|
||||
defaultValue = 0.0
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return float number decoded from ``value``."""
|
||||
return struct.unpack("<q", value)[0] / 10000.
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return string containing encoded ``value``."""
|
||||
return struct.pack("<q", round(value * 10000))
|
||||
|
||||
class DbfLogicalFieldDef(DbfFieldDef):
|
||||
"""Definition of the logical field."""
|
||||
|
||||
typeCode = "L"
|
||||
defaultValue = -1
|
||||
length = 1
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return True, False or -1 decoded from ``value``."""
|
||||
# Note: value always is 1-char string
|
||||
if value == "?":
|
||||
return -1
|
||||
if value in "NnFf ":
|
||||
return False
|
||||
if value in "YyTt":
|
||||
return True
|
||||
raise ValueError("[%s] Invalid logical value %r" % (self.name, value))
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a character from the "TF?" set.
|
||||
|
||||
Return:
|
||||
Return value is "T" if ``value`` is True
|
||||
"?" if value is -1 or False otherwise.
|
||||
|
||||
"""
|
||||
if value is True:
|
||||
return "T"
|
||||
if value == -1:
|
||||
return "?"
|
||||
return "F"
|
||||
|
||||
|
||||
class DbfMemoFieldDef(DbfFieldDef):
|
||||
"""Definition of the memo field.
|
||||
|
||||
Note: memos aren't currenly completely supported.
|
||||
|
||||
"""
|
||||
|
||||
typeCode = "M"
|
||||
defaultValue = " " * 10
|
||||
length = 10
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return int .dbt block number decoded from the string object."""
|
||||
#return int(value)
|
||||
raise NotImplementedError
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return raw data string encoded from a ``value``.
|
||||
|
||||
Note: this is an internal method.
|
||||
|
||||
"""
|
||||
#return str(value)[:self.length].ljust(self.length)
|
||||
raise NotImplementedError
|
||||
|
||||
|
||||
class DbfDateFieldDef(DbfFieldDef):
|
||||
"""Definition of the date field."""
|
||||
|
||||
typeCode = "D"
|
||||
defaultValue = utils.classproperty(lambda cls: datetime.date.today())
|
||||
# "yyyymmdd" gives us 8 characters
|
||||
length = 8
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a ``datetime.date`` instance decoded from ``value``."""
|
||||
if value.strip():
|
||||
return utils.getDate(value)
|
||||
else:
|
||||
return None
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a string-encoded value.
|
||||
|
||||
``value`` argument should be a value suitable for the
|
||||
`utils.getDate` call.
|
||||
|
||||
Return:
|
||||
Return value is a string in format "yyyymmdd".
|
||||
|
||||
"""
|
||||
if value:
|
||||
return utils.getDate(value).strftime("%Y%m%d")
|
||||
else:
|
||||
return " " * self.length
|
||||
|
||||
|
||||
class DbfDateTimeFieldDef(DbfFieldDef):
|
||||
"""Definition of the timestamp field."""
|
||||
|
||||
# a difference between JDN (Julian Day Number)
|
||||
# and GDN (Gregorian Day Number). note, that GDN < JDN
|
||||
JDN_GDN_DIFF = 1721425
|
||||
typeCode = "T"
|
||||
defaultValue = utils.classproperty(lambda cls: datetime.datetime.now())
|
||||
# two 32-bits integers representing JDN and amount of
|
||||
# milliseconds respectively gives us 8 bytes.
|
||||
# note, that values must be encoded in LE byteorder.
|
||||
length = 8
|
||||
|
||||
def decodeValue(self, value):
|
||||
"""Return a `datetime.datetime` instance."""
|
||||
assert len(value) == self.length
|
||||
# LE byteorder
|
||||
_jdn, _msecs = struct.unpack("<2I", value)
|
||||
if _jdn >= 1:
|
||||
_rv = datetime.datetime.fromordinal(_jdn - self.JDN_GDN_DIFF)
|
||||
_rv += datetime.timedelta(0, _msecs / 1000.0)
|
||||
else:
|
||||
# empty date
|
||||
_rv = None
|
||||
return _rv
|
||||
|
||||
def encodeValue(self, value):
|
||||
"""Return a string-encoded ``value``."""
|
||||
if value:
|
||||
value = utils.getDateTime(value)
|
||||
# LE byteorder
|
||||
_rv = struct.pack("<2I", value.toordinal() + self.JDN_GDN_DIFF,
|
||||
(value.hour * 3600 + value.minute * 60 + value.second) * 1000)
|
||||
else:
|
||||
_rv = "\0" * self.length
|
||||
assert len(_rv) == self.length
|
||||
return _rv
|
||||
|
||||
|
||||
_fieldsRegistry = {}
|
||||
|
||||
def registerField(fieldCls):
|
||||
"""Register field definition class.
|
||||
|
||||
``fieldCls`` should be subclass of the `DbfFieldDef`.
|
||||
|
||||
Use `lookupFor` to retrieve field definition class
|
||||
by the type code.
|
||||
|
||||
"""
|
||||
assert fieldCls.typeCode is not None, "Type code isn't defined"
|
||||
# XXX: use fieldCls.typeCode.upper()? in case of any decign
|
||||
# don't forget to look to the same comment in ``lookupFor`` method
|
||||
_fieldsRegistry[fieldCls.typeCode] = fieldCls
|
||||
|
||||
|
||||
def lookupFor(typeCode):
|
||||
"""Return field definition class for the given type code.
|
||||
|
||||
``typeCode`` must be a single character. That type should be
|
||||
previously registered.
|
||||
|
||||
Use `registerField` to register new field class.
|
||||
|
||||
Return:
|
||||
Return value is a subclass of the `DbfFieldDef`.
|
||||
|
||||
"""
|
||||
# XXX: use typeCode.upper()? in case of any decign don't
|
||||
# forget to look to the same comment in ``registerField``
|
||||
return _fieldsRegistry[chr(typeCode)]
|
||||
|
||||
## register generic types
|
||||
|
||||
for (_name, _val) in list(globals().items()):
|
||||
if isinstance(_val, type) and issubclass(_val, DbfFieldDef) \
|
||||
and (_name != "DbfFieldDef"):
|
||||
__all__.append(_name)
|
||||
registerField(_val)
|
||||
del _name, _val
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,273 @@
|
||||
"""DBF header definition.
|
||||
|
||||
TODO:
|
||||
- handle encoding of the character fields
|
||||
(encoding information stored in the DBF header)
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
16-sep-2010 [als] fromStream: fix century of the last update field
|
||||
11-feb-2007 [als] added .ignoreErrors
|
||||
10-feb-2007 [als] added __getitem__: return field definitions
|
||||
by field name or field number (zero-based)
|
||||
04-jul-2006 [als] added export declaration
|
||||
15-dec-2005 [yc] created
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.6 $"[11:-2]
|
||||
__date__ = "$Date: 2010/09/16 05:06:39 $"[7:-2]
|
||||
|
||||
__all__ = ["DbfHeader"]
|
||||
|
||||
import io
|
||||
import datetime
|
||||
import struct
|
||||
import time
|
||||
import sys
|
||||
|
||||
from . import fields
|
||||
from .utils import getDate
|
||||
|
||||
|
||||
class DbfHeader(object):
|
||||
"""Dbf header definition.
|
||||
|
||||
For more information about dbf header format visit
|
||||
`http://www.clicketyclick.dk/databases/xbase/format/dbf.html#DBF_STRUCT`
|
||||
|
||||
Examples:
|
||||
Create an empty dbf header and add some field definitions:
|
||||
dbfh = DbfHeader()
|
||||
dbfh.addField(("name", "C", 10))
|
||||
dbfh.addField(("date", "D"))
|
||||
dbfh.addField(DbfNumericFieldDef("price", 5, 2))
|
||||
Create a dbf header with field definitions:
|
||||
dbfh = DbfHeader([
|
||||
("name", "C", 10),
|
||||
("date", "D"),
|
||||
DbfNumericFieldDef("price", 5, 2),
|
||||
])
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = ("signature", "fields", "lastUpdate", "recordLength",
|
||||
"recordCount", "headerLength", "changed", "_ignore_errors")
|
||||
|
||||
## instance construction and initialization methods
|
||||
|
||||
def __init__(self, fields=None, headerLength=0, recordLength=0,
|
||||
recordCount=0, signature=0x03, lastUpdate=None, ignoreErrors=False,
|
||||
):
|
||||
"""Initialize instance.
|
||||
|
||||
Arguments:
|
||||
fields:
|
||||
a list of field definitions;
|
||||
recordLength:
|
||||
size of the records;
|
||||
headerLength:
|
||||
size of the header;
|
||||
recordCount:
|
||||
number of records stored in DBF;
|
||||
signature:
|
||||
version number (aka signature). using 0x03 as a default meaning
|
||||
"File without DBT". for more information about this field visit
|
||||
``http://www.clicketyclick.dk/databases/xbase/format/dbf.html#DBF_NOTE_1_TARGET``
|
||||
lastUpdate:
|
||||
date of the DBF's update. this could be a string ('yymmdd' or
|
||||
'yyyymmdd'), timestamp (int or float), datetime/date value,
|
||||
a sequence (assuming (yyyy, mm, dd, ...)) or an object having
|
||||
callable ``ticks`` field.
|
||||
ignoreErrors:
|
||||
error processing mode for DBF fields (boolean)
|
||||
|
||||
"""
|
||||
self.signature = signature
|
||||
if fields is None:
|
||||
self.fields = []
|
||||
else:
|
||||
self.fields = list(fields)
|
||||
self.lastUpdate = getDate(lastUpdate)
|
||||
self.recordLength = recordLength
|
||||
self.headerLength = headerLength
|
||||
self.recordCount = recordCount
|
||||
self.ignoreErrors = ignoreErrors
|
||||
# XXX: I'm not sure this is safe to
|
||||
# initialize `self.changed` in this way
|
||||
self.changed = bool(self.fields)
|
||||
|
||||
# @classmethod
|
||||
def fromString(cls, string):
|
||||
"""Return header instance from the string object."""
|
||||
return cls.fromStream(io.StringIO(str(string)))
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
# @classmethod
|
||||
def fromStream(cls, stream):
|
||||
"""Return header object from the stream."""
|
||||
stream.seek(0)
|
||||
first_32 = stream.read(32)
|
||||
if type(first_32) != bytes:
|
||||
_data = bytes(first_32, sys.getfilesystemencoding())
|
||||
_data = first_32
|
||||
(_cnt, _hdrLen, _recLen) = struct.unpack("<I2H", _data[4:12])
|
||||
#reserved = _data[12:32]
|
||||
_year = _data[1]
|
||||
if _year < 80:
|
||||
# dBase II started at 1980. It is quite unlikely
|
||||
# that actual last update date is before that year.
|
||||
_year += 2000
|
||||
else:
|
||||
_year += 1900
|
||||
## create header object
|
||||
_obj = cls(None, _hdrLen, _recLen, _cnt, _data[0],
|
||||
(_year, _data[2], _data[3]))
|
||||
## append field definitions
|
||||
# position 0 is for the deletion flag
|
||||
_pos = 1
|
||||
_data = stream.read(1)
|
||||
while _data != b'\r':
|
||||
_data += stream.read(31)
|
||||
_fld = fields.lookupFor(_data[11]).fromString(_data, _pos)
|
||||
_obj._addField(_fld)
|
||||
_pos = _fld.end
|
||||
_data = stream.read(1)
|
||||
return _obj
|
||||
fromStream = classmethod(fromStream)
|
||||
|
||||
## properties
|
||||
|
||||
year = property(lambda self: self.lastUpdate.year)
|
||||
month = property(lambda self: self.lastUpdate.month)
|
||||
day = property(lambda self: self.lastUpdate.day)
|
||||
|
||||
def ignoreErrors(self, value):
|
||||
"""Update `ignoreErrors` flag on self and all fields"""
|
||||
self._ignore_errors = value = bool(value)
|
||||
for _field in self.fields:
|
||||
_field.ignoreErrors = value
|
||||
ignoreErrors = property(
|
||||
lambda self: self._ignore_errors,
|
||||
ignoreErrors,
|
||||
doc="""Error processing mode for DBF field value conversion
|
||||
|
||||
if set, failing field value conversion will return
|
||||
``INVALID_VALUE`` instead of raising conversion error.
|
||||
|
||||
""")
|
||||
|
||||
## object representation
|
||||
|
||||
def __repr__(self):
|
||||
_rv = """\
|
||||
Version (signature): 0x%02x
|
||||
Last update: %s
|
||||
Header length: %d
|
||||
Record length: %d
|
||||
Record count: %d
|
||||
FieldName Type Len Dec
|
||||
""" % (self.signature, self.lastUpdate, self.headerLength,
|
||||
self.recordLength, self.recordCount)
|
||||
_rv += "\n".join(
|
||||
["%10s %4s %3s %3s" % _fld.fieldInfo() for _fld in self.fields]
|
||||
)
|
||||
return _rv
|
||||
|
||||
## internal methods
|
||||
|
||||
def _addField(self, *defs):
|
||||
"""Internal variant of the `addField` method.
|
||||
|
||||
This method doesn't set `self.changed` field to True.
|
||||
|
||||
Return value is a length of the appended records.
|
||||
Note: this method doesn't modify ``recordLength`` and
|
||||
``headerLength`` fields. Use `addField` instead of this
|
||||
method if you don't exactly know what you're doing.
|
||||
|
||||
"""
|
||||
# insure we have dbf.DbfFieldDef instances first (instantiation
|
||||
# from the tuple could raise an error, in such a case I don't
|
||||
# wanna add any of the definitions -- all will be ignored)
|
||||
_defs = []
|
||||
_recordLength = 0
|
||||
for _def in defs:
|
||||
if isinstance(_def, fields.DbfFieldDef):
|
||||
_obj = _def
|
||||
else:
|
||||
(_name, _type, _len, _dec) = (tuple(_def) + (None,) * 4)[:4]
|
||||
_cls = fields.lookupFor(_type)
|
||||
_obj = _cls(_name, _len, _dec,
|
||||
ignoreErrors=self._ignore_errors)
|
||||
_recordLength += _obj.length
|
||||
_defs.append(_obj)
|
||||
# and now extend field definitions and
|
||||
# update record length
|
||||
self.fields += _defs
|
||||
return _recordLength
|
||||
|
||||
## interface methods
|
||||
|
||||
def addField(self, *defs):
|
||||
"""Add field definition to the header.
|
||||
|
||||
Examples:
|
||||
dbfh.addField(
|
||||
("name", "C", 20),
|
||||
dbf.DbfCharacterFieldDef("surname", 20),
|
||||
dbf.DbfDateFieldDef("birthdate"),
|
||||
("member", "L"),
|
||||
)
|
||||
dbfh.addField(("price", "N", 5, 2))
|
||||
dbfh.addField(dbf.DbfNumericFieldDef("origprice", 5, 2))
|
||||
|
||||
"""
|
||||
_oldLen = self.recordLength
|
||||
self.recordLength += self._addField(*defs)
|
||||
if not _oldLen:
|
||||
self.recordLength += 1
|
||||
# XXX: may be just use:
|
||||
# self.recordeLength += self._addField(*defs) + bool(not _oldLen)
|
||||
# recalculate headerLength
|
||||
self.headerLength = 32 + (32 * len(self.fields)) + 1
|
||||
self.changed = True
|
||||
|
||||
def write(self, stream):
|
||||
"""Encode and write header to the stream."""
|
||||
stream.seek(0)
|
||||
stream.write(self.toString())
|
||||
fields = [_fld.toString() for _fld in self.fields]
|
||||
stream.write(''.join(fields).encode(sys.getfilesystemencoding()))
|
||||
stream.write(b'\x0D') # cr at end of all header data
|
||||
self.changed = False
|
||||
|
||||
def toString(self):
|
||||
"""Returned 32 chars length string with encoded header."""
|
||||
return struct.pack("<4BI2H",
|
||||
self.signature,
|
||||
self.year - 1900,
|
||||
self.month,
|
||||
self.day,
|
||||
self.recordCount,
|
||||
self.headerLength,
|
||||
self.recordLength) + (b'\x00' * 20)
|
||||
#TODO: figure out if bytes(utf-8) is correct here.
|
||||
|
||||
def setCurrentDate(self):
|
||||
"""Update ``self.lastUpdate`` field with current date value."""
|
||||
self.lastUpdate = datetime.date.today()
|
||||
|
||||
def __getitem__(self, item):
|
||||
"""Return a field definition by numeric index or name string"""
|
||||
if isinstance(item, str):
|
||||
_name = item.upper()
|
||||
for _field in self.fields:
|
||||
if _field.name == _name:
|
||||
return _field
|
||||
else:
|
||||
raise KeyError(item)
|
||||
else:
|
||||
# item must be field index
|
||||
return self.fields[item]
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,266 @@
|
||||
"""DBF record definition.
|
||||
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] __repr__: added special case for invalid field values
|
||||
10-feb-2007 [als] added .rawFromStream()
|
||||
30-oct-2006 [als] fix record length in .fromStream()
|
||||
04-jul-2006 [als] added export declaration
|
||||
20-dec-2005 [yc] DbfRecord.write() -> DbfRecord._write();
|
||||
added delete() method.
|
||||
16-dec-2005 [yc] record definition moved from `dbf`.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.7 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 09:05:49 $"[7:-2]
|
||||
|
||||
__all__ = ["DbfRecord"]
|
||||
|
||||
import sys
|
||||
|
||||
from . import utils
|
||||
|
||||
class DbfRecord(object):
|
||||
"""DBF record.
|
||||
|
||||
Instances of this class shouldn't be created manualy,
|
||||
use `dbf.Dbf.newRecord` instead.
|
||||
|
||||
Class implements mapping/sequence interface, so
|
||||
fields could be accessed via their names or indexes
|
||||
(names is a preffered way to access fields).
|
||||
|
||||
Hint:
|
||||
Use `store` method to save modified record.
|
||||
|
||||
Examples:
|
||||
Add new record to the database:
|
||||
db = Dbf(filename)
|
||||
rec = db.newRecord()
|
||||
rec["FIELD1"] = value1
|
||||
rec["FIELD2"] = value2
|
||||
rec.store()
|
||||
Or the same, but modify existed
|
||||
(second in this case) record:
|
||||
db = Dbf(filename)
|
||||
rec = db[2]
|
||||
rec["FIELD1"] = value1
|
||||
rec["FIELD2"] = value2
|
||||
rec.store()
|
||||
|
||||
"""
|
||||
|
||||
__slots__ = "dbf", "index", "deleted", "fieldData"
|
||||
|
||||
## creation and initialization
|
||||
|
||||
def __init__(self, dbf, index=None, deleted=False, data=None):
|
||||
"""Instance initialiation.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance this record belonogs to.
|
||||
index:
|
||||
An integer record index or None. If this value is
|
||||
None, record will be appended to the DBF.
|
||||
deleted:
|
||||
Boolean flag indicating whether this record
|
||||
is a deleted record.
|
||||
data:
|
||||
A sequence or None. This is a data of the fields.
|
||||
If this argument is None, default values will be used.
|
||||
|
||||
"""
|
||||
self.dbf = dbf
|
||||
# XXX: I'm not sure ``index`` is necessary
|
||||
self.index = index
|
||||
self.deleted = deleted
|
||||
if data is None:
|
||||
self.fieldData = [_fd.defaultValue for _fd in dbf.header.fields]
|
||||
else:
|
||||
self.fieldData = list(data)
|
||||
|
||||
# XXX: validate self.index before calculating position?
|
||||
position = property(lambda self: self.dbf.header.headerLength + \
|
||||
self.index * self.dbf.header.recordLength)
|
||||
|
||||
def rawFromStream(cls, dbf, index):
|
||||
"""Return raw record contents read from the stream.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance containing the record.
|
||||
index:
|
||||
Index of the record in the records' container.
|
||||
This argument can't be None in this call.
|
||||
|
||||
Return value is a string containing record data in DBF format.
|
||||
|
||||
"""
|
||||
# XXX: may be write smth assuming, that current stream
|
||||
# position is the required one? it could save some
|
||||
# time required to calculate where to seek in the file
|
||||
dbf.stream.seek(dbf.header.headerLength +
|
||||
index * dbf.header.recordLength)
|
||||
return dbf.stream.read(dbf.header.recordLength)
|
||||
rawFromStream = classmethod(rawFromStream)
|
||||
|
||||
def fromStream(cls, dbf, index):
|
||||
"""Return a record read from the stream.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance new record should belong to.
|
||||
index:
|
||||
Index of the record in the records' container.
|
||||
This argument can't be None in this call.
|
||||
|
||||
Return value is an instance of the current class.
|
||||
|
||||
"""
|
||||
return cls.fromString(dbf, cls.rawFromStream(dbf, index), index)
|
||||
fromStream = classmethod(fromStream)
|
||||
|
||||
def fromString(cls, dbf, string, index=None):
|
||||
"""Return record read from the string object.
|
||||
|
||||
Arguments:
|
||||
dbf:
|
||||
A `Dbf.Dbf` instance new record should belong to.
|
||||
string:
|
||||
A string new record should be created from.
|
||||
index:
|
||||
Index of the record in the container. If this
|
||||
argument is None, record will be appended.
|
||||
|
||||
Return value is an instance of the current class.
|
||||
|
||||
"""
|
||||
return cls(dbf, index, string[0]=="*",
|
||||
[_fd.decodeFromRecord(string) for _fd in dbf.header.fields])
|
||||
fromString = classmethod(fromString)
|
||||
|
||||
## object representation
|
||||
|
||||
def __repr__(self):
|
||||
_template = "%%%ds: %%s (%%s)" % max([len(_fld)
|
||||
for _fld in self.dbf.fieldNames])
|
||||
_rv = []
|
||||
for _fld in self.dbf.fieldNames:
|
||||
_val = self[_fld]
|
||||
if _val is utils.INVALID_VALUE:
|
||||
_rv.append(_template %
|
||||
(_fld, "None", "value cannot be decoded"))
|
||||
else:
|
||||
_rv.append(_template % (_fld, _val, type(_val)))
|
||||
return "\n".join(_rv)
|
||||
|
||||
## protected methods
|
||||
|
||||
def _write(self):
|
||||
"""Write data to the dbf stream.
|
||||
|
||||
Note:
|
||||
This isn't a public method, it's better to
|
||||
use 'store' instead publically.
|
||||
Be design ``_write`` method should be called
|
||||
only from the `Dbf` instance.
|
||||
|
||||
|
||||
"""
|
||||
self._validateIndex(False)
|
||||
self.dbf.stream.seek(self.position)
|
||||
self.dbf.stream.write(bytes(self.toString(),
|
||||
sys.getfilesystemencoding()))
|
||||
# FIXME: may be move this write somewhere else?
|
||||
# why we should check this condition for each record?
|
||||
if self.index == len(self.dbf):
|
||||
# this is the last record,
|
||||
# we should write SUB (ASCII 26)
|
||||
self.dbf.stream.write(b"\x1A")
|
||||
|
||||
## utility methods
|
||||
|
||||
def _validateIndex(self, allowUndefined=True, checkRange=False):
|
||||
"""Valid ``self.index`` value.
|
||||
|
||||
If ``allowUndefined`` argument is True functions does nothing
|
||||
in case of ``self.index`` pointing to None object.
|
||||
|
||||
"""
|
||||
if self.index is None:
|
||||
if not allowUndefined:
|
||||
raise ValueError("Index is undefined")
|
||||
elif self.index < 0:
|
||||
raise ValueError("Index can't be negative (%s)" % self.index)
|
||||
elif checkRange and self.index <= self.dbf.header.recordCount:
|
||||
raise ValueError("There are only %d records in the DBF" %
|
||||
self.dbf.header.recordCount)
|
||||
|
||||
## interface methods
|
||||
|
||||
def store(self):
|
||||
"""Store current record in the DBF.
|
||||
|
||||
If ``self.index`` is None, this record will be appended to the
|
||||
records of the DBF this records belongs to; or replaced otherwise.
|
||||
|
||||
"""
|
||||
self._validateIndex()
|
||||
if self.index is None:
|
||||
self.index = len(self.dbf)
|
||||
self.dbf.append(self)
|
||||
else:
|
||||
self.dbf[self.index] = self
|
||||
|
||||
def delete(self):
|
||||
"""Mark method as deleted."""
|
||||
self.deleted = True
|
||||
|
||||
def toString(self):
|
||||
"""Return string packed record values."""
|
||||
# for (_def, _dat) in zip(self.dbf.header.fields, self.fieldData):
|
||||
#
|
||||
|
||||
return "".join([" *"[self.deleted]] + [
|
||||
_def.encodeValue(_dat)
|
||||
for (_def, _dat) in zip(self.dbf.header.fields, self.fieldData)
|
||||
])
|
||||
|
||||
def asList(self):
|
||||
"""Return a flat list of fields.
|
||||
|
||||
Note:
|
||||
Change of the list's values won't change
|
||||
real values stored in this object.
|
||||
|
||||
"""
|
||||
return self.fieldData[:]
|
||||
|
||||
def asDict(self):
|
||||
"""Return a dictionary of fields.
|
||||
|
||||
Note:
|
||||
Change of the dicts's values won't change
|
||||
real values stored in this object.
|
||||
|
||||
"""
|
||||
return dict([_i for _i in zip(self.dbf.fieldNames, self.fieldData)])
|
||||
|
||||
def __getitem__(self, key):
|
||||
"""Return value by field name or field index."""
|
||||
if isinstance(key, int):
|
||||
# integer index of the field
|
||||
return self.fieldData[key]
|
||||
# assuming string field name
|
||||
return self.fieldData[self.dbf.indexOfFieldName(key)]
|
||||
|
||||
def __setitem__(self, key, value):
|
||||
"""Set field value by integer index of the field or string name."""
|
||||
if isinstance(key, int):
|
||||
# integer index of the field
|
||||
return self.fieldData[key]
|
||||
# assuming string field name
|
||||
self.fieldData[self.dbf.indexOfFieldName(key)] = value
|
||||
|
||||
# vim: et sts=4 sw=4 :
|
||||
@@ -0,0 +1,170 @@
|
||||
"""String utilities.
|
||||
|
||||
TODO:
|
||||
- allow strings in getDateTime routine;
|
||||
"""
|
||||
"""History (most recent first):
|
||||
11-feb-2007 [als] added INVALID_VALUE
|
||||
10-feb-2007 [als] allow date strings padded with spaces instead of zeroes
|
||||
20-dec-2005 [yc] handle long objects in getDate/getDateTime
|
||||
16-dec-2005 [yc] created from ``strutil`` module.
|
||||
"""
|
||||
|
||||
__version__ = "$Revision: 1.4 $"[11:-2]
|
||||
__date__ = "$Date: 2007/02/11 08:57:17 $"[7:-2]
|
||||
|
||||
import datetime
|
||||
import time
|
||||
|
||||
|
||||
def unzfill(str):
|
||||
"""Return a string without ASCII NULs.
|
||||
|
||||
This function searchers for the first NUL (ASCII 0) occurance
|
||||
and truncates string till that position.
|
||||
|
||||
"""
|
||||
try:
|
||||
return str[:str.index(b'\0')]
|
||||
except ValueError:
|
||||
return str
|
||||
|
||||
|
||||
def getDate(date=None):
|
||||
"""Return `datetime.date` instance.
|
||||
|
||||
Type of the ``date`` argument could be one of the following:
|
||||
None:
|
||||
use current date value;
|
||||
datetime.date:
|
||||
this value will be returned;
|
||||
datetime.datetime:
|
||||
the result of the date.date() will be returned;
|
||||
string:
|
||||
assuming "%Y%m%d" or "%y%m%dd" format;
|
||||
number:
|
||||
assuming it's a timestamp (returned for example
|
||||
by the time.time() call;
|
||||
sequence:
|
||||
assuming (year, month, day, ...) sequence;
|
||||
|
||||
Additionaly, if ``date`` has callable ``ticks`` attribute,
|
||||
it will be used and result of the called would be treated
|
||||
as a timestamp value.
|
||||
|
||||
"""
|
||||
if date is None:
|
||||
# use current value
|
||||
return datetime.date.today()
|
||||
if isinstance(date, datetime.date):
|
||||
return date
|
||||
if isinstance(date, datetime.datetime):
|
||||
return date.date()
|
||||
if isinstance(date, (int, float)):
|
||||
# date is a timestamp
|
||||
return datetime.date.fromtimestamp(date)
|
||||
if isinstance(date, str):
|
||||
date = date.replace(" ", "0")
|
||||
if len(date) == 6:
|
||||
# yymmdd
|
||||
return datetime.date(*time.strptime(date, "%y%m%d")[:3])
|
||||
# yyyymmdd
|
||||
return datetime.date(*time.strptime(date, "%Y%m%d")[:3])
|
||||
if hasattr(date, "__getitem__"):
|
||||
# a sequence (assuming date/time tuple)
|
||||
return datetime.date(*date[:3])
|
||||
return datetime.date.fromtimestamp(date.ticks())
|
||||
|
||||
|
||||
def getDateTime(value=None):
|
||||
"""Return `datetime.datetime` instance.
|
||||
|
||||
Type of the ``value`` argument could be one of the following:
|
||||
None:
|
||||
use current date value;
|
||||
datetime.date:
|
||||
result will be converted to the `datetime.datetime` instance
|
||||
using midnight;
|
||||
datetime.datetime:
|
||||
``value`` will be returned as is;
|
||||
string:
|
||||
*** CURRENTLY NOT SUPPORTED ***;
|
||||
number:
|
||||
assuming it's a timestamp (returned for example
|
||||
by the time.time() call;
|
||||
sequence:
|
||||
assuming (year, month, day, ...) sequence;
|
||||
|
||||
Additionaly, if ``value`` has callable ``ticks`` attribute,
|
||||
it will be used and result of the called would be treated
|
||||
as a timestamp value.
|
||||
|
||||
"""
|
||||
if value is None:
|
||||
# use current value
|
||||
return datetime.datetime.today()
|
||||
if isinstance(value, datetime.datetime):
|
||||
return value
|
||||
if isinstance(value, datetime.date):
|
||||
return datetime.datetime.fromordinal(value.toordinal())
|
||||
if isinstance(value, (int, float)):
|
||||
# value is a timestamp
|
||||
return datetime.datetime.fromtimestamp(value)
|
||||
if isinstance(value, str):
|
||||
raise NotImplementedError("Strings aren't currently implemented")
|
||||
if hasattr(value, "__getitem__"):
|
||||
# a sequence (assuming date/time tuple)
|
||||
return datetime.datetime(*tuple(value)[:6])
|
||||
return datetime.datetime.fromtimestamp(value.ticks())
|
||||
|
||||
|
||||
class classproperty(property):
|
||||
"""Works in the same way as a ``property``, but for the classes."""
|
||||
|
||||
def __get__(self, obj, cls):
|
||||
return self.fget(cls)
|
||||
|
||||
|
||||
class _InvalidValue(object):
|
||||
|
||||
"""Value returned from DBF records when field validation fails
|
||||
|
||||
The value is not equal to anything except for itself
|
||||
and equal to all empty values: None, 0, empty string etc.
|
||||
In other words, invalid value is equal to None and not equal
|
||||
to None at the same time.
|
||||
|
||||
This value yields zero upon explicit conversion to a number type,
|
||||
empty string for string types, and False for boolean.
|
||||
|
||||
"""
|
||||
|
||||
def __eq__(self, other):
|
||||
return not other
|
||||
|
||||
def __ne__(self, other):
|
||||
return not (other is self)
|
||||
|
||||
def __bool__(self):
|
||||
return False
|
||||
|
||||
def __int__(self):
|
||||
return 0
|
||||
__long__ = __int__
|
||||
|
||||
def __float__(self):
|
||||
return 0.0
|
||||
|
||||
def __str__(self):
|
||||
return ""
|
||||
|
||||
def __unicode__(self):
|
||||
return ""
|
||||
|
||||
def __repr__(self):
|
||||
return "<INVALID>"
|
||||
|
||||
# invalid value is a constant singleton
|
||||
INVALID_VALUE = _InvalidValue()
|
||||
|
||||
# vim: set et sts=4 sw=4 :
|
||||
+102
@@ -450,6 +450,108 @@ class TablibTestCase(unittest.TestCase):
|
||||
self.assertEqual(_tsv, data.tsv)
|
||||
|
||||
|
||||
def test_dbf_import_set(self):
|
||||
data.append(self.john)
|
||||
data.append(self.george)
|
||||
data.headers = self.headers
|
||||
|
||||
_dbf = data.dbf
|
||||
data.dbf = _dbf
|
||||
|
||||
#self.assertEqual(_dbf, data.dbf)
|
||||
try:
|
||||
self.assertEqual(_dbf, data.dbf)
|
||||
except AssertionError:
|
||||
index = 0
|
||||
so_far = ''
|
||||
for reg_char, data_char in zip(_dbf, data.dbf):
|
||||
so_far += chr(data_char)
|
||||
if reg_char != data_char and index not in [1, 2, 3]:
|
||||
raise AssertionError('Failing at char %s: %s vs %s %s' % (
|
||||
index, reg_char, data_char, so_far))
|
||||
index += 1
|
||||
|
||||
def test_dbf_export_set(self):
|
||||
"""Test DBF import."""
|
||||
data.append(self.john)
|
||||
data.append(self.george)
|
||||
data.append(self.tom)
|
||||
data.headers = self.headers
|
||||
|
||||
_regression_dbf = (b'\x03r\x06\x06\x03\x00\x00\x00\x81\x00\xab\x00\x00'
|
||||
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00\x00FIRST_NAME\x00C\x00\x00\x00\x00P\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00LAST_NAME\x00\x00C\x00'
|
||||
b'\x00\x00\x00P\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00GPA\x00\x00\x00\x00\x00\x00\x00\x00N\x00\x00\x00\x00\n'
|
||||
b'\x08\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\r'
|
||||
)
|
||||
_regression_dbf += b' John' + (b' ' * 75)
|
||||
_regression_dbf += b' Adams' + (b' ' * 74)
|
||||
_regression_dbf += b' 90.0000000'
|
||||
_regression_dbf += b' George' + (b' ' * 73)
|
||||
_regression_dbf += b' Washington' + (b' ' * 69)
|
||||
_regression_dbf += b' 67.0000000'
|
||||
_regression_dbf += b' Thomas' + (b' ' * 73)
|
||||
_regression_dbf += b' Jefferson' + (b' ' * 70)
|
||||
_regression_dbf += b' 50.0000000'
|
||||
_regression_dbf += b'\x1a'
|
||||
|
||||
if is_py3:
|
||||
# If in python3, decode regression string to binary.
|
||||
#_regression_dbf = bytes(_regression_dbf, 'utf-8')
|
||||
#_regression_dbf = _regression_dbf.replace(b'\n', b'\r')
|
||||
pass
|
||||
|
||||
try:
|
||||
self.assertEqual(_regression_dbf, data.dbf)
|
||||
except AssertionError:
|
||||
index = 0
|
||||
found_so_far = ''
|
||||
for reg_char, data_char in zip(_regression_dbf, data.dbf):
|
||||
#found_so_far += chr(data_char)
|
||||
if reg_char != data_char and index not in [1, 2, 3]:
|
||||
raise AssertionError(
|
||||
'Failing at char %s: %s vs %s (found %s)' % (
|
||||
index, reg_char, data_char, found_so_far))
|
||||
index += 1
|
||||
|
||||
def test_dbf_format_detect(self):
|
||||
"""Test the DBF format detection."""
|
||||
_dbf = (b'\x03r\x06\x03\x03\x00\x00\x00\x81\x00\xab\x00\x00'
|
||||
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00\x00FIRST_NAME\x00C\x00\x00\x00\x00P\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00LAST_NAME\x00\x00C\x00'
|
||||
b'\x00\x00\x00P\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
|
||||
b'\x00\x00GPA\x00\x00\x00\x00\x00\x00\x00\x00N\x00\x00\x00\x00\n'
|
||||
b'\x08\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\r'
|
||||
)
|
||||
_dbf += b' John' + (b' ' * 75)
|
||||
_dbf += b' Adams' + (b' ' * 74)
|
||||
_dbf += b' 90.0000000'
|
||||
_dbf += b' George' + (b' ' * 73)
|
||||
_dbf += b' Washington' + (b' ' * 69)
|
||||
_dbf += b' 67.0000000'
|
||||
_dbf += b' Thomas' + (b' ' * 73)
|
||||
_dbf += b' Jefferson' + (b' ' * 70)
|
||||
_dbf += b' 50.0000000'
|
||||
_dbf += b'\x1a'
|
||||
|
||||
_yaml = '- {age: 90, first_name: John, last_name: Adams}'
|
||||
_tsv = 'foo\tbar'
|
||||
_csv = '1,2,3\n4,5,6\n7,8,9\n'
|
||||
_json = '[{"last_name": "Adams","age": 90,"first_name": "John"}]'
|
||||
|
||||
_bunk = (
|
||||
'¡¡¡¡¡¡¡¡£™∞¢£§∞§¶•¶ª∞¶•ªº••ª–º§•†•§º¶•†¥ª–º•§ƒø¥¨©πƒø†ˆ¥ç©¨√øˆ¥≈†ƒ¥ç©ø¨çˆ¥ƒçø¶'
|
||||
)
|
||||
self.assertTrue(tablib.formats.dbf.detect(_dbf))
|
||||
self.assertFalse(tablib.formats.dbf.detect(_yaml))
|
||||
self.assertFalse(tablib.formats.dbf.detect(_tsv))
|
||||
self.assertFalse(tablib.formats.dbf.detect(_csv))
|
||||
self.assertFalse(tablib.formats.dbf.detect(_json))
|
||||
self.assertFalse(tablib.formats.dbf.detect(_bunk))
|
||||
|
||||
def test_csv_format_detect(self):
|
||||
"""Test CSV format detection."""
|
||||
|
||||
|
||||
Reference in New Issue
Block a user