Pandas Interpolate Does not Handle Empty Values at Front of Series

[cancan101]

This code means the interpolator won't handle extrapolating to the front of the Series even though the underlying implementations may have no problem with the extrapolation.

See for example UnivariateSpline whose default behavior is extrapolation.

Interpolation works fine at the end of the Series.

For example:

s = pd.Series([1, 2, 3, 4, np.nan, 6, np.nan])
s.interpolate(method='spline', order=1)
0    1
1    2
2    3
3    4
4    5
5    6
6    7
dtype: float64

but:

s = pd.Series([np.nan, 2, 3, 4, np.nan, 6, 7])
s.interpolate(method='spline', order=1)
0   NaN
1     2
2     3
3     4
4     5
5     6
6     7
dtype: float64

[jreback]

hmm, I suppose that if the underlying can handle it then would be ok to change this. Try it and see if anything blows up. You might need to expand the tests a bit to make sure that this is all tested.

[lmjohns3]

This seems related to #9218 -- the only reason the values aren't filled at the beginning of the DataFrame/Series is that the fill for interpolate only happens forward.

[cancan101]

Sure. This lines in the referenced ticket seems to sum it up:

The limit refers to forward filling only, even though interpolation is not inherently directional.

[lmjohns3]

Agreed.

Also see #10691 for an ongoing pull request/discussion.

[github-repo-health-uic]

This code means the interpolator won't handle extrapolating to the front of the Series even though the underlying implementations may have no problem with the extrapolation.

See for example UnivariateSpline whose default behavior is extrapolation.

Interpolation works fine at the end of the Series.

For example:

s = pd.Series([1, 2, 3, 4, np.nan, 6, np.nan])
s.interpolate(method='spline', order=1)
0    1
1    2
2    3
3    4
4    5
5    6
6    7
dtype: float64

but:

s = pd.Series([np.nan, 2, 3, 4, np.nan, 6, 7])
s.interpolate(method='spline', order=1)
0   NaN
1     2
2     3
3     4
4     5
5     6
6     7
dtype: float64

You can set limit_direction='both' . This is a parameter in the interpolation()