df.loc column where name contains "x" and count rows where rows contains "y"-CodePudding

I have a pandas df and want to loc columns containing string "X" and then sum for every row where rows string contains "Y".

Desired results

ABC	ABC	ABX	ABX	Result
A	A	YA	A	1
A	A	YA	YA	2

I've tried beginning below, ,but already here and error occurs.

df['result'] = np.where(df.loc[:~df.columns.str.contains('X')]

Error: TypeError: cannot do slice indexing on DatetimeIndex with these indexers

Anyone know how to go about?

CodePudding user response：

You can use filter to select the columns and stack str.contains groupby.sum to count the values:

df['Result'] = (df.filter(like='X')
                  .stack()
                  .str.contains('Y')
                  .groupby(level=0).sum()
               )

Or with apply:

df['Result'] = (df.filter(like='X')
                  .apply(lambda s: s.str.contains('Y'))
                  .sum(axis=1)
               )

output:

  ABC ABC ABX ABX  Result
0   A   A  YA   A       1
1   A   A  YA  YA       2