I have a DataFrame

import pandas as pd
import numpy as np

index = pd.MultiIndex.from_product([["A", "B"], ["AA", "BB"]])
columns = pd.MultiIndex.from_product([["X", "Y"], ["XX", "YY"]])

df = pd.DataFrame([[1,2,3,4],
                   [5,6,7,8],
                   [9,10,11,12],
                   [13,14,15,16]], index = index, columns = columns)

and slice

toSkip = ((slice(None), slice(None)), (["X"], slice(None)))

I know that I can write df.loc[slice] to get the subset of DataFrame which corresponds to this slice. But how can I do the opposite so get the difference between original df and the one obtained with that slice?

CodePudding user response：

How to invert slicing

To get the idea let's make it more complicated.

import pandas as pd
import numpy as np

index = pd.MultiIndex.from_product([["A", "B", "C"], ["AA", "BB", "CC"]])
columns = pd.MultiIndex.from_product([["X", "Y", "Z"], ["XX", "YY", "ZZ"]])
data = (
    np
    .arange(len(index) * len(columns))
    .reshape(len(index), len(columns))
)

df = pd.DataFrame(data, index, columns)

Let's say I want to process all the data except the inner square (B,Y).

I can get the square by slicing. To get others I'm gonna use a boolean mask:

mask = pd.DataFrame(True, index, columns)
toSkip = ((['B'], slice(None)), (['Y'], slice(None)))
mask.loc[toSkip] = False

Now I can transform others by windowing with mask:

# just for illustration purposes
# let's invert the sign of numbers
df[mask] *= -1

Here's the output:

CodePudding user response：

If slice is a Series with boolean values, then logical negation operator ~ will give the opposite of the condition. So,

df[~slice]

will return rows that doesn't satisfy the condition slice

CodePudding user response：

Not sure if this is you want, you can drop the index and columns of toSkip dataframe

toSkip = ((slice(None), slice(None)), (["X"], slice(None)))

tmp = df.loc[toSkip]
out = df.drop(index=tmp.index, columns=tmp.columns)

print(out)

Empty DataFrame
Columns: [(Y, XX), (Y, YY)]
Index: []