DatetimeIndexes with day frequency hang when they have more than 2 elements

[cpcloud]

See #3609. Copied (more or less) from there

dind = period_range('1/1/2001', '1/1/2002').to_timestamp()
s = Series(randn(dind.size), dind)
s.resample('T', kind='period')  # hangs here

Doesn't hang (throws an error) for the simple case of

dind = period_range('1/1/2001', '1/2/2001').to_timestamp()
s = Series(randn(dind.size), dind)
s.resample('T', kind='period')

and starts to hang for dind.size > 2.

[cpcloud]

problem is an object conversion of a 526000-ish size array. large PeriodIndexes have a very slow astype method due to iterating over each element in python...might help to cythonize but not sure how common this is. in any case should probably raise or cythonize

[jreback]

@cpcloud I think should do this for 0.13. I can do if you don't want.

[jreback]

k gr8

[jreback]

hows this coming?

[cpcloud]

pr will close #3899

[jreback]

gr8

[cpcloud]

@jreback i was slightly mistaken ... this ended up being unrelated to #3899 ... i'll still fix this one tho

[cpcloud]

related #1688

[cpcloud]

@jreback what do u want to do here? the hanging is caused by construction of a giant period index....not really much there except move the construction down into cython...should i do that? or just close as won't fix

[jreback]

you mean like this issue: #5155 (ffor constructing giant periodindex is broken but easily fixed)

but isn't this issue different?

[cpcloud]

no that's the same issue

[cpcloud]

it's not that more than two elements causes it it's that the construction of a huge period index (bc going from years to minutes) makes it seems as if it's hanging

[cpcloud]

closing in favor of #5155

[jreback]

ahh...ok...makes sense