DataFrame.applymap with identity function changes dtype

[craustin]

from datetime import datetime
from pandas import DataFrame
df = DataFrame({'x1': [datetime(1996,1,1)]})
df2 = df.applymap(lambda x: x)

Issue:
type(df['x1'][0]) == datetime.datetime
type(df2['x1'][0]) == pandas.lib.Timestamp

This repros in the 11/27/2012 cut of 0.10.0 - but not in 0.9.0 release.

[wesm]

Is there somewhere where Timestamp is causing you problems? It's a subclass of datetime.datetime and should be interchangeable

[craustin]

Here's an unexpected crash:
from datetime import datetime
from pandas import DataFrame
from pandas.core.datetools import bday
df = DataFrame({'x1': [datetime(1996,1,1)]})
df = df.applymap(lambda x: x+bday)
df = df.applymap(lambda x: x+bday)

The first applymap succeeds (but alters the dtypes as in the above post). The second applymap raises this exception:

pandas\core\frame.pyc in applymap(self, func)
4114 applied : DataFrame
4115 """
-> 4116 return self.apply(lambda x: lib.map_infer(x, func))
4117
4118 #----------------------------------------------------------------------

pandas\core\frame.pyc in apply(self, func, axis, broadcast, raw, args, **kwds)
3981 return self._apply_raw(f, axis)
3982 else:
-> 3983 return self._apply_standard(f, axis)
3984 else:
3985 return self._apply_broadcast(f, axis)

pandas\core\frame.pyc in _apply_standard(self,
func, axis, ignore_failures)
4056 # no k defined yet
4057 pass
-> 4058 raise e
4059
4060 if len(results) > 0 and _is_sequence(results[0]):

TypeError: ("ufunc 'add' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule 'safe'", u'occurred at index x1')

[wesm]

I see the issue. Will get a fix in for 0.10

[wesm]

All set now

[craustin]

Great. Thanks Wes.