Dbfread

^^^^^^^^^^^^^^^^^^

* bugfix: crashed when record list was not terminated with b'\x1a'.
(Bug first apperad in 1.0.2 after a rewrite.)

* bugfix: memo fields with no value were returned as ''. They are
now returned correctly as None.

* bugfix: field header terminaters were compared with strings.

* added example parserclass_debugstring.py.

^^^^^^^^^^^^^^^^^^

* reinstated hastily removed parserclass option.

^^^^^^^^^^^^^^^^^^

* added example record_objects.py.

* removed parserclass option to allow for internal changes. There is
currently no (documented) way to add custom field types.

^^^^^^^^^^^^^^^^^^

* bugfix: deleted records were ignored when using open().

* memo file is now opened and closed by each iterator instead of
staying open all the time.

^^^^^^^^^^^^^^^^^^

* records can now be streamed from the file, making it possible to
read data files that are too large to fit in memory.

* documentation is more readable and complete.

* now installs correctly with easy_install.

* added "--encoding" option to dbf2sqlite which can be used to
override character encoding.

^^^^^^^^^^^^^^^^^^

Initial release.


API Changes
===========

``dbfread.open()`` and ``dbfread.read()`` are deprecated as of version
``2.0``, and will be removed in ``2.2``.

The ``DBF`` class is no longer a subclass of ``list``. This makes the
API a lot cleaner and easier to understand, but old code that relied
on this behaviour will be broken. Iteration and record counting works
the same as before. Other list operations can be rewritten using the
``record`` attribute. For example::

table = dbfread.read('people.dbf')
print(table[1])

can be rewritten as::

table = DBF('people.dbf', load=True)
print(table.records[1])

``open()`` and ``read()`` both return ``DeprecatedDBF``, which is a
subclass of ``DBF`` and ``list`` and thus backward compatible.