Pandas MultiIndex Lookup By Equality and Set Membership

Question

Given a pandas Series, or Dataframe, with a multiindex:

first_key = ['a', 'b', 'c']
second_key = [1, 2, 3]

m_index = pd.MultiIndex.from_product([first_key, second_key],
                                     names=['first_key', 'second_key'])

series_with_index = pd.Series(0.0, index=m_index)

How can the MultiIndex be indexed to lookup an equality for the first level and an isin on the second index?

For example, how can all values where the first level is equal to a and the second level is in the set {2, 3, 4} be set to 1.0?

Thank you in advance for your consideration and response.

Asked By: Ramón J Romero y Vigil

||

Source

Answer 1

Try this:

you can use index.get_level_values() to find all the values in the first level that equal a.

index.isin() has a level parameter, so you can pass your set into that.

lastly change the values in the series to 1 where they are both True

m1 = series_with_index.index.get_level_values(0) == 'a'
m2 = series_with_index.index.isin({2,3,4},level=1)

series_with_index.mask(m1 & m2,1)

Output:

first_key  second_key
a          1             0.0
           2             1.0
           3             1.0
b          1             0.0
           2             0.0
           3             0.0
c          1             0.0
           2             0.0
           3             0.0

Answered By: rhug123

Answer 2

With composite conditions for each of 2 levels:

s = pd.Series(0.0, index=m_index)
s[(s.index.get_level_values(0) == 'a') & ((s.index.get_level_values(1).isin({2, 3, 4})))] = 1.0

first_key  second_key
a          1             0.0
           2             1.0
           3             1.0
b          1             0.0
           2             0.0
           3             0.0
c          1             0.0
           2             0.0
           3             0.0

Answered By: RomanPerekhrest

Answer 3

You can use get_level_values and perform boolean indexing:

m1 = series_with_index.index.get_level_values('first_key')=='a'
m2 = series_with_index.index.get_level_values('second_key').isin([2,3,4])
series_with_index[m1&m2] = 1

Updated Series:

first_key  second_key
a          1             0.0
           2             1.0
           3             1.0
b          1             0.0
           2             0.0
           3             0.0
c          1             0.0
           2             0.0
           3             0.0
dtype: float64

Answered By: mozway

Answer 4

query works with named index, but only available for DataFrame:

my_set = {2, 3, 4}
series_with_index.to_frame()
    .query('first_key=="a" & second_key in @my_set')
    .iloc[:,0]

Answered By: Quang Hoang

Pandas MultiIndex Lookup By Equality and Set Membership

Question:

Answers: