API for selecting ranged groups

[cpcloud]

I often find myself writing code that looks like this:

start_group = df[df.str_col == 'start'].index.values + 1
end_group = df[df.str_col == 'end'].index.values - 1
for s, e in zip(start_group, end_group):
    some_function(df[s:e])

to get frame subsets between repeated start and end indicators from a particular column.

Possible API:

df.grange(marker_col='str_col', start='start', end='end')

which would return an iterator over the the subsets.

I'm not sure if this is general enough to include in library code. I'll happily close the issue if it's not. There also might be a way to do this already.

[jreback]

kind of like #3685

and like split #4059, make add this example there?

[dsm054]

Requests for improved split/partition tools now make up, what, a half-dozen or so enhancement requests? (Including one of mine. :^)

[jreback]

@cpcloud can you find all of the split/partition issues and put them in one master issue (maybe checkboxes?), possibly creating another tag (maybe split?) instead of using groupby tag?

[cpcloud]

sure thing

[jreback]

@cpcloud can you consolidate these issues when you have a chance (putting in a single master isse) with checkboxes