Pandas returns 0/y=inf, even when y is a non-zero dataframe or series.

[JohnNapier]

Pandas returns the wrong result for the operation 0/y where y is a dataframe or series of non-zero values:

import pandas as pd
y=pd.Series(range(1,10))
print 0/y # wrong

However, this operation would work:

for i in y:print 0/i # correct

I'm using Pandas 0.15.2-1.

[jreback]

this is a dupe of #8445

a pull-request to fix is welcome

[JohnNapier]

@keszybz, @jreback, I don't think this is a dupe. In the other issue, the answer should be NaN. In this issue, the answer should be zero. Do you agree?

If so, I suggest to re-open.

[jreback]

It exactly the same causation. A bug in fill_zeros; I cross referenced this, to make this a tests as well.

[JohnNapier]

@jreback, adding as a test is good. Thanks.

[jreback]

pull-requests are welcome!

This actually is a very straightforward change (meaning I think I put a pointer to exactly where it should be changed and its only a single function). The logic is just off in a case I think.

hint hint hint :)