ANN: 0.94.003

Ethan Furman ethan at
Thu Jul 26 21:31:01 CEST 2012

A few more bug fixes, and I actually included the documentation this 
time.  :)  It can be found at, and has been 
tested on CPythons 2.4 - 2.7, and PyPy 1.8.

dbf v0.94.003

dbf (also known as python dbase) is a module for reading/writing
dBase III, FP, VFP, and Clipper .dbf database files.  It's
an ancient format that still finds lots of use (the most common
I'm aware of is retrieving legacy data so it can be stored in a
newer database system; other uses include GIS, stand-alone programs
such as Family History, Personal Finance, etc.).


Table -- represents a single .dbf/.dbt (or .fpt) file combination
and provides access to records; suports the sequence access and 'with'
protocols.  Temporary tables can also live entirely in memory.

Record -- repesents a single record/row in the table, with field access
returning native or custom data types; supports the sequence, mapping,
attribute access (with the field names as the attributes), and 'with'
protocols.  Updates to a record object are reflected on disk either
immediately (using gather() or write()), or at the end of a 'with'

     dBase III (Null not supported)

         Character --> unicode
         Date      --> or None
         Logical   --> bool or None
         Memo      --> unicode or None
         Numeric   --> int/float depending on field definition or None

         Float     --> same as numeric

     Clipper (Null not supported)

         Character --> unicode  (character fields can be up to 65,519)

     Foxpro (Null supported)

         General   --> str (treated as binary)
         Picture   --> str (treated as binary)

     Visual Foxpro (Null supported)

         Currency  --> decimal.Decimal
         douBle    --> float
         Integer   --> int
         dateTime  --> datetime.datetime

     If a field is uninitialized (Date, Logical, Numeric, Memo, General,
     Picture) then None is returned for the value.

Custom data types:

     Null     -->  used to support Null values

     Char     -->  unicode type that auto-trims trailing whitespace, and
                   ignores trailing whitespace for comparisons

     Date     -->  date object that allows for no date

     DateTime -->  datetime object that allows for no datetime

     Time     -->  time object that allows for no time

     Logical  -->  adds Unknown state to bool's: instead of True/False/None,
                   values are Truth, Falsth, and Unknown, with appropriate
                   tri-state logic; while bool(None) is False, bool(Unknown)
                   raises a TypeError;  __index__ of Unknown is 2

     Quantum  -->  similar to Logical, but implements boolean algebra (I 

Whirlwind Tour

import datetime
import dbf

table = dbf.Table(':test:', 'name C(25); age N(3,0); birth D; qualified L')

for datum in (
         ('Spanky', 7, dbf.Date.fromymd('20010315'), False),
         ('Spunky', 23, dbf.Date(1989, 07, 23), True),
         ('Sparky', 99, dbf.Date(), dbf.Unknown),

for record in table:
     print record
     print '--------'
     print record[0:3]
     print record['name':'birth']
     print [, record.age, record.birth]
     print '--------'

custom =
         default_data_types=dict(C=dbf.Char, D=dbf.Date, L=dbf.Logical),

with custom:    # automatically opened and closed
     for record in table:
     for record in custom:
         print record
         print '--------'
         print record[0:3]
         print record['name':'birth']
         print [, record.age, record.birth]
         print '--------'


More information about the Python-list mailing list