CODEWITHSUNDEEP
pythonpandas
Alexander Ivanov
Alexander Ivanov

Reputation: 325

Can't use loc with DatetimeIndex

I can't select using loc when there is DatetimeIndex.

test = pd.DataFrame(data=np.array([[0, 0], [0, 2], [1, 3]), columns=pd.date_range(start='2019-01-01', end='2019-01-02', freq='D'))
test.loc[test>1, '2019-01-02']

I expect it to return pandas.Series([2, 3]), but it returns the error "Cannot index with multidimensional key"

Upvotes: 1

Views: 189

Answers (1)

aiguofer
aiguofer

Reputation: 2135

In this case, your index is not a DatetimeIndex, only your columns are. The issue is that when you use test>1 as a comparison, it will return a DataFrame with the same size as test with Booleans for each cell showing whether the value is > 1. When you pass an array of booleans, it expects it to be a 1 dimensional array, but since you're passing it a DataFrame (2 dimensional), you get the "multidemensional key" error. I believe what you want here is:

test.loc[test['2019-01-02']>1, '2019-01-02']

Upvotes: 1

Related Questions