I'm using this to conditionally select rows of column:
X.loc[data['column'] == 1]
But I want to expand this condition to several columns. These columns have something in common: They contain a same string. So actually I have a column1, a column2, ... , column100 etc. and this condition should apply to all of these columns. Actually something like this (wildcard):
X.loc[data['column*'] == 1]
These conditions should be linked with OR. Any chance to do this easily?
For some dataframe X
p A p B p C
0 0 0 0
1 0 0 0
2 0 0 1
3 0 0 0
4 0 0 0
5 0 0 0
6 1 0 0
If you can set up the names of the columns you want to test for in col_list
col_list = X.columns
You can then use np.any() to test with or between each:
X.loc[(X[col_list] == 1).any(axis=1)]
Which gives you:
p A p B p C
2 0 0 1
6 1 0 0
Informed you don't need loc and will still get the same answer, credit to @MaartynFabre for the info
X[(X[col_list] == 1).any(axis=1)]
p A p B p C
2 0 0 1
6 1 0 0
col0 col1 col2
0 1 1 2
1 1 1 1
2 2 2 2
make a new dataframe with the test for all columns
result_s = d.concat((df['col%i'%i] == 1 for i in range(3)), axis=1).all(axis=1)
results in
0 False
1 True
2 False
dtype: bool
if you do df[result_s] you get
col0 col1 col2
1 1 1 1
this selects the rows where all columns are ==1 If one of the is enough, change the .all() to .any
col0 col1 col2
0 1 1 2
1 1 1 1
columnOne, columnTwo etc.
df[(df[selected_columns]==1).any(axis=1)] can work too and be more efficientdf['col%i'%i] == 1 for i in range(3) is just a method to select the columns, depending on how they look this might need modification. You will need to provide more information in your question to make this clearPut each comparison in brackets and combine them with logical operators:
pd.DataFrame(X).loc[(data['col1']==23) & (data['col2']==42)] # and
pd.DataFrame(X).loc[(data['col1']==23) | (data['col2']==42)] # or
Here's another way to consider:
df
col0 col1 col2
0 1 1 2
1 1 1 1
2 2 2 2
df.loc[df['col0'] == 1, [x for x in df.columns if x == 'col0']]
col0
0 1
1 1
You can use list comprehension to find the columns you're looking for.
numpy.any()Two questions I asked very similar to yours: stackoverflow.com/questions/43677505/… and stackoverflow.com/questions/42647710/…col_list, TryX.loc[(X[col_list] == 1).any(axis=1)].locis not needed