CODEWITHSUNDEEP
pythonpandasdataframemulti-index
user77005
user77005

Reputation: 1859

pandas dataframe check if index exists in a multi index

I have a pandas Dataframe which has a multiindex created using the columns userid and itemid. df looks like this 

                  0     1     2
userid  itemid
007     5000      9     4     3
007     4000      6     7     1
009     3000      1     2     3

I want to check if the index [007, 6000] exists in the dataframe df. How can I do that. If I run the following code there is an error TypeError: unhashable type: 'list'. 

if [007, 6000] in df.index:
    print('it works')

Upvotes: 14

Views: 25292

Answers (3)

Markus Dutschke
Markus Dutschke

Reputation: 10606

cast pd.MultiIndex to list and check for presence in list

code

import pandas as pd

mi = pd.MultiIndex.from_tuples(
    [(7, 5000), (7, 4000), (8, 3000)], names=['usedId', 'itemId'])
df = pd.DataFrame([[9, 4, 3], [6, 7, 1], [1, 2, 3]], index=mi)

print('df:', df, sep='\n', end='\n\n')
print('mi:', mi, sep='\n', end='\n\n')

print('Check for elements in Multi-Index:')
print('\t(7, 4000) in mi.to_list():', (7, 4000) in mi.to_list())
print('\t(7, 99) in mi.to_list():', (7, 99) in mi.to_list())

output

df:
               0  1  2
usedId itemId         
7      5000    9  4  3
       4000    6  7  1
8      3000    1  2  3

mi:
MultiIndex([(7, 5000),
            (7, 4000),
            (8, 3000)],
           names=['usedId', 'itemId'])

Check for elements in Multi-Index:
    (7, 4000) in mi.to_list(): True
    (7, 99) in mi.to_list(): False

Upvotes: 0

jezrael
jezrael

Reputation: 863256

Use Index.isin:

df = df.index.isin([('007','5000')])
print (df)
[ True False False]

Upvotes: 5

cs95
cs95

Reputation: 402852

For this - 

df

               0  1  2
userid itemid         
7      5000    9  4  3
       4000    6  7  1
9      3000    1  2  3

df.index.values
array([(7, 5000), (7, 4000), (9, 3000)], dtype=object)

You can use df.index.isin.

df.index.isin([(7, 5000)])
array([ True, False, False], dtype=bool)

This gives you a mask corresponding to where that value can be found. If you just want to know whether it exists or not, use np.ndarray.any in conjunction with isin.

df.index.isin([(7, 5000)]).any()
True

df.index.isin([(7, 6000)]).any()
False

Upvotes: 20

Related Questions