CODEWITHSUNDEEP
pythonpandasfilter
KingOtto
KingOtto

Reputation: 1493

Get multiple columns of DataFrame by filter (pandas.filter `like` for multiple strings)

What is the easist way to retrieve all columns from a DataFrame based on a part of their name, but with different selection patterns? I am too lazy to type the exact column names, and I am looking for a convenient way.

Say, I want all columns that start with the code 'T1*' or 'T2*' - ideally a command like

df = pd.DataFrame([[1, 2, 3], [4, 5, 6], [7, 8, 9]], columns={'T101', 'T201', 'G101'})
df.filter(like=['T1', 'T2'])

which is not supported, as like='' only takes 1 string.

Slow workaround I currently use:

col_list = df.columns
target_cols = [e for e in col_list if any(se in e for se in ['T1','T2'])]
df[target_cols]

which gives me the desired output:

>>> df[target_cols]
   T201  T101
0     2     3
1     5     6
2     8     9

Is there any faster/more convenient way?

Note this is not about the actual data in the rows/I am not looking for a mask to select rows.

Upvotes: 0

Views: 127

Answers (1)

Anurag Dabas
Anurag Dabas

Reputation: 24314

you can use regex parameter in filter() method:

df.filter(regex='^T1|^T2')

output:

   T101     T201
0   1       3
1   4       6
2   7       9

Upvotes: 2

Related Questions