ENH: Updating a dataframe counting the cells that will be updated.

[ccsv]

Counting the cells that were updated when an update is performed to a dataframe.

Currently there are no commands to count the number of cells being updated so you can put it in a log or something when df.update(df2) is performed.

See:
http://stackoverflow.com/questions/23102757/python-pandas-updating-dataframe-and-counting-the-number-of-cells-updated

[danielballan]

It's unclear how this information should be returned to the user. Can you suggest an example of how the usage would work?

Offhand, I think the user should just compare the two DataFrames to get that information. The little-used df.align() might come in handy here.

[jreback]

I suppose could add an argument like
return_counts

but seems a very narrow use case

[cpcloud]

@ccsv Can you give a bit more detail about your use case?

[ccsv]

@jreback @cpcloud @danielballan
I would say either a list of cells that were changed or counts.
The reason for this is if you are working with multiple people and someone made an accidental update you can ID data to 'back out' or erase. I can use this function to create a log of my changes.

[jtratner]

This would add a lot of complexity to the internals for, arguably, little gain. If we were to do this, I'd suggest adding a separate method that takes in two DataFrames and returns a description of the differences between them. It's unlikely that we'd produce anything more memory efficient anyways.
That said, I'm not sure there's value in providing some sugar over a few lines of code.

[jreback]

a cookbook entry would be nice for this