Pandas return index and column name based on condition-CodePudding

Let's say we have a DataFrame:

df = pd.DataFrame({
    'D': {'A': 0.810032, 'B': 0.752299, 'C': 0.224038},
    'E': {'A': 0.17714, 'B': 0.680405, 'C': 0.769732},
    'F': {'A': 0.942959, 'B': 0.436582, 'C': 0.269791}
})
print(df)
          D         E         F
A  0.810032  0.177140  0.942959
B  0.752299  0.680405  0.436582
C  0.224038  0.769732  0.269791

Is it possible to return index and column names into a list of tuples based on a condition, for instance

print(df < 0.25)
   D      E      F
A  False   True  False
B  False  False  False
C  True  False  False

I would like to end up with a list of tuples like:

[('A','E'),('C','D')]

CodePudding user response：

One option is to reshape the data and filter with the booleans:

df.lt(.25).stack().loc[lambda df: df].index.to_list()

[('A', 'E'), ('C', 'D')]

CodePudding user response：

We can also use DataFrame.where then stack which takes advantage of the fact that stack drops NaN values by default then use Index.tolist to get the list of tuples:

results = df.where(df < .25).stack().index.tolist()

results:

[('A', 'E'), ('C', 'D')]