CODEWITHSUNDEEP
pythonindexingpandasslice
dyingoptimist
dyingoptimist

Reputation: 113

strange behaviour in Pandas slicing

Find it little strange that slicing with -ve number and a list made out of that slice aren't matching. Why is it so?

dfl = DataFrame(np.random.randn(5,2),columns=list('AB'))

>>> dfl
          A         B
0 -0.775699 -0.622863
1  2.418822  0.747961
2  1.000536  0.036125
3  0.394238 -0.215857
4  0.374322  0.795460

>>> dfl.iloc[:,-1:]
          B
0 -0.622863
1  0.747961
2  0.036125
3 -0.215857
4  0.795460

>>> list(dfl.iloc[:,-1:])
['B']


>>> list(dfl.iloc[:,1])
[-0.622863, 0.747961, 0.036125, -0.215857, 0.795460]

Upvotes: 1

Views: 91

Answers (1)

EdChum
EdChum

Reputation: 394031

The reason is because your first slicing returns a DataFrame whilst the latter returns a Series:

In [183]:

type(dfl.iloc[:,-1:])
Out[183]:
pandas.core.frame.DataFrame

In [184]:

type(dfl.iloc[:,1])
Out[184]:
pandas.core.series.Series

Calling list on a df returns the columns

In [189]:

list(dfl)
Out[189]:
['A', 'B']

Upvotes: 1

Related Questions