EHN: DataFrame.rolling().rank()

[Lingou Deng]

I would like to propose a rolling rank computation for dataframe, as I cannot find a efficient version so far.

a rank for [1,1.5,5,5,8,nan,nan,2,3,5] should be[0.125,0.25,0.75,0.75,1,nan,nan,0.375,0.5,0.75], and the rolling rank should 0.75 which is the value for last position.

I think it should be adding value to "Rolling" class. I would be happy to contribute in the codebase.

Any thoughts?

Lingou

[Jeff Reback]

We have an issue for that https://github.com/pydata/pandas/issues/9481.

Contributions certainly welcome!

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.