arrays used as indices must be of integer or boolean type

Package:

pandas

30911

Exception Class:

IndexError

Raise code

try:
            indexer = np.asarray(indexer, dtype=np.intp)
        except ValueError as err:
            raise ValueError(
                "Cannot index with an integer indexer containing NA values"
            ) from err
    else:
        raise IndexError("arrays used as indices must be of integer or boolean type")

    return indexer

Links to the raise (1)

https://github.com/pandas-dev/pandas/blob/b3e335254f46a526ee3ce9bb757eac4011d9d1fe/pandas/core/indexers.py#L571

Ways to fix

Error code:

import pandas as pd 

indexer = pd.array(['one','two'])
arr = pd.array([1, 2, 3])
pd.api.indexers.check_array_indexer(arr, indexer)

check_array_indexer checks if indexer is a valid array indexer for array. Take 2 arguments: an array that is being indexed and an indexer that is used to index. The error comes when the indexer has a different data type than integer and boolean.

Fix code:

import pandas as pd 

indexer = pd.array([0,2])
arr = pd.array([1, 2, 3])
pd.api.indexers.check_array_indexer(arr, indexer)

Jul 20, 2021

anonim 13.0k

Add a possible fix

Please authorize to post fix