allowing datetime and timedelta datatype in pd cut bins

[aileronajay]

xref #14714, follow-on to #14737

[aileronajay]

The change is currently WIP, will add tests and other change, @jorisvandenbossche

[codecov-io]

Current coverage is 84.64% (diff: 88.88%)

Merging #14798 into master will increase coverage by <.01%

@@             master     #14798   diff @@
==========================================
  Files           144        144          
  Lines         51021      51030     +9   
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
+ Hits          43188      43196     +8   
- Misses         7833       7834     +1   
  Partials          0          0          

Powered by Codecov. Last update f79bc7a...82bffa1

[jreback]

add the issue number as a comment

[aileronajay]

@jreback i dont think there is an open issue for this change, @jorisvandenbossche had proposed this change when i was making changes to cut to allow datetime and timedelta data types

[jreback]

use pd.Timestamp(....) instead of direct np.datetime64

[jreback]

actually you prob want to test with datetime.datetime, Timestamp, np.datetime64 (but just put them in a loop something like

data = ['2012-12-12', '2012-12-14']

for conv in [Timestamp(x).to_pydatetime, Timestamp, np.datetime64]:
      bins = [ conv(v) for v in data ]

also test
bins = pd.to_datetime(data)

these should all work, because internally you need to wrap a Timestamp converter around each of fhe bins (if dtype==M8) or Timedelta if dtype==m8

[aileronajay]

@jreback made the changes, i added a new method in tile.py to handle the time type bins

[jorisvandenbossche]

@aileronajay can you update this?

[aileronajay]

@jorisvandenbossche i am caught up with some stuff right now, i should be able to make these changes next week

[jreback]

thanks!