CODEWITHSUNDEEP
python-3.xpandaspandas-groupby
dark horse
dark horse

Reputation: 3719

Pandad - KeyError : Column not in index

I have a pandas Dataframe that has a list of columns in a groupby function. I get an error

KeyError : "['Type1'] not in index"

Given below is the code that throws the error

temp_v1 = temp_df.groupby(level, as_index = False).sum()[[level, 'Type1', 'Type2','Type3', 'Type4', 'Type5']]

Could anyone guide me where am I going wrong with the above Dataframe. Thanks..

Upvotes: 0

Views: 1380

Answers (1)

jezrael
jezrael

Reputation: 862661

I guess problem is string column Type1:

level = 'F'
temp_df = pd.DataFrame({
         'Type1':list('abcdef'),
         'Type2':[4,5,4,5,5,4],
         'Type3':[7,8,9,4,2,3],
         'Type4':[1,3,5,7,1,0],
         'Type5':[5,3,6,9,2,4],
         'col':[5,3,6,9,2,4],
         'F':list('aaabbb')
})

print (temp_df.dtypes)
Type1    object
Type2     int64
Type3     int64
Type4     int64
Type5     int64
col       int64
F        object
dtype: object

Solution is add list before sum function, but Type1 column is excluded, because not numeric.

cols = [level, 'Type1', 'Type2','Type3', 'Type4', 'Type5']
temp_v1 = temp_df.groupby(level, as_index = False)[cols].sum()
print (temp_v1)
   F  Type2  Type3  Type4  Type5
0  a     13     24      9     14
1  b     14      9      8     15

Another problem is typo in column name or traling whitespace, you can check it by converting columns names to list:

print (temp_df.columns.tolist())
['Type1', 'Type2', 'Type3', 'Type4', 'Type5', 'col', 'F']

Upvotes: 2

Related Questions