dbfpy

bug: if you can't write dbf file with IOError, use "rb+" : self.stream = file(f, ("r+b", "rb+")[bool(readOnly)]) enhance: you can save str to num(int or float) cells Fields.py------------------------- class DbfNumericFieldDef(DbfFieldDef): ..... def encodeValue(self, value): """Return string containing encoded ``value``.""" if isinstance(value, str): #print str(value)[:self.length].ljust(self.length) return str(value)[:self.length].rjust(self.length) else: _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 .....

Just what we needed to parse the DBF part of some shapefiles.

I found a bug ni you code.Like the following codes: from dbfpy import dbf db = dbf.Dbf("test.dbf", new=True) db.addField( ("BIRTHDATE", "D"), ) rec = db.newRecord() #year=1000 is before 1900; the datetime strftime() methods require year >= 1900 rec["BIRTHDATE"] = "10001010" rec.store() db.close() thanks

python 2.6 ok. has minor problems with Y2K+ dates. fix proposed (and can be applied on per-user basis) -- see project's tracker.