Pandas ENH: allow transform to handle differently indexed outputs

from SO

using transform to broadcast the result works fine if the index is an Int64Index because the input and output are the same. However, if you actually need the index (e.g. in this example we are doing an .at_time(...) to lookup a particular value, then broadcasting fails.

I think the way to do this is to define .at_time as a transform function to handle this case.

Comment From: mroeschke

Guess there hasn't been a lot of interest or traction for this feature, and generally since pandas is getting stricter with apply like functionality probably best to have the indexes align so closing