Apply a map to DataFrame rows which are NaN

Question

I want to apply a mapping to a Pandas DataFrame, where some rows in one specific column are NaN. These are left overs from a prior mapping.

mymap = defaultdict(str)
mymap["a"] = "-1 test"
mymap["b"] = "-2 test"
mymap["c"] = "-3 test"

df[ df["my_infos"].isnull() ] = df["something"].map(lambda ip: map_function(ip, mymap))

Here is the function:

def map_function(ip, mymap):
    # do stuff
    for key, value in mymap.iteritems():
           # do stuff
           return stuff
    return other_stuff

This terminates after it iterates over the whole column, but shows:

File "/opt/anaconda2/lib/python2.7/site-packages/pandas/core/indexing.py", line 527, in _setitem_with_indexer raise ValueError('Must have equal len keys and value ' ValueError: Must have equal len keys and value when setting with an iterable

Is this the correct way to select all rows in the my_infos column, which are NaN? I somehow sense... that's not the case.

Answer 1

I think you can select NaN values to both sides and then map:

df[ df["my_infos"].isnull() ] = 
df.ix[ df["my_infos"].isnull(), "something"].map(lambda ip: map_function(ip, mymap))