How to get a series from a pandas dataframe using a series of column names?

Question

I have a pandas dataframe df with numeric data. I also have a series s with the same index as df and values consisting of df column labels, e.g.

import pandas as pd
df = pd.DataFrame(
    index=[0, 1, 2], columns=[0, 1, 2],
    data=[[1, 2, 3], [4, 5, 6], [7, 8, 9]]
)
s = pd.Series(index=[0, 1, 2]), data=[0, 1, 2])

How can I use s to slice df and get another series s1 that contains df values corresponding to the (index, value) pairs in s as their .loc() identifiers in df, i.e.

s1 = pd.Series(index=[0, 1, 2], data=[1, 5, 9])

Answer 1

Use DataFrame.lookup to lookup the values in df based on s.index and s, then create a new series from this lookup values:

s1 = pd.Series(df.lookup(s.index, s), index=s.index)

Another idea using DataFrame.stack and indexing using DataFrame.loc:

s1 = df.stack().loc[zip(s.index, s)].droplevel(1)

Result:

print(s1)

0    1
1    5
2    9
dtype: int64