Mapping with unicode strings buggy the format operator %

[Martin von Loewis]

chemacortes at wanadoo.WIPE_ME.es (Chema Cortes) writes:

> I have a dictionary with unicode strings as keys. Formating it with the
> format operator % raise errors when the key cannot convert into a
> non-unicode string. I try to surround the error, but the only thing I
> can do is not to use unicode keys in dictionaries. I supose that bug is
> caused because the format operator use the re module, not the sre (with
> unicode support).

No, it's a different problem: for %(name)s arguments in Unicode
strings, name is converted to UTF-8; the resulting UTF-8 string is
then searched in the dictionary.

> >>> d={ "año":2001 }
> >>> "%(año)4d" % d
> '2001'
> >>> d={ u"año":2001 }
> >>> "%(año)4d"  % d      # KeyError: año

This fails because u"año" and "año" cannot be compared: byte strings
and Unicode strings are comparable only if the byte string is ASCII.

> >>> u"%(año)4d" % d      # KeyError: año

This fails because the key is converted to UTF-8 before lookup.

I recommend to report a bug to sf.net/projects/python.

Regards,
Martin