CODEWITHSUNDEEP
pythonpandas
littlely
littlely

Reputation: 1428

transform method of pandas can not pass multi methods

I want to pass 2 methods to transform method of pandas as the API says it can pass a list of functions or dict of column names -> functions. I pass a list of functions, but it does not work:

import pandas as pd
import numpy as np
df = pd.DataFrame({'rti':['a','a','b','c','b','c','a'],'ts':[10,10,9,12,9,13,11],'rs':[8,8,22,11,12,11,9]})
df.groupby('rti').transform(['mean','sum'])

It shows:"TypeError: unhashable type: 'list'"

Upvotes: 1

Views: 117

Answers (2)

Alexander
Alexander

Reputation: 109546

It works, but the functions used must not be aggregation functions such as sum, max or min.

>>> df.transform([np.abs, np.sign])
        rs            ts     
  absolute sign absolute sign
0        8    1       10    1
1        8    1       10    1
2       22    1        9    1
3       11    1       12    1
4       12    1        9    1
5       11    1       13    1
6        9    1       11    1

Refer to the documentation here. Note that the transform method for groupby objects accepts only a function (not a list of functions which is for the dataframe transform method).

Per the doc string of the tranform method of groupby objects:

Signature: gb.transform(func, *args, **kwargs)

Docstring: Call function producing a like-indexed DataFrame on each group and return a DataFrame having the same indexes as the original object filled with the transformed values

Parameters

f : function Function to apply to each group

Notes

Each group is endowed the attribute 'name' in case you need to know which group you are working on.

The current implementation imposes three requirements on f:

  • f must return a value that either has the same shape as the input subframe or can be broadcast to the shape of the input subframe. For example, f returns a scalar it will be broadcast to have the same shape as the input subframe.
  • if this is a DataFrame, f must support application column-by-column in the subframe. If f also supports application to the entire subframe, then a fast path is used starting from the second chunk.
  • f must not mutate groups. Mutation is not supported and may produce unexpected results.

Upvotes: 1

BENY
BENY

Reputation: 323226

Seems like transform do not accept list of function , Open issue in github 

df.groupby('rti').agg(['mean','sum']).reindex(df.rti)
Out[12]: 
            rs             ts    
          mean sum       mean sum
rti                              
a     8.333333  25  10.333333  31
a     8.333333  25  10.333333  31
b    17.000000  34   9.000000  18
c    11.000000  22  12.500000  25
b    17.000000  34   9.000000  18
c    11.000000  22  12.500000  25
a     8.333333  25  10.333333  31

Upvotes: 2

Related Questions