Add first and last method to IndexedFrame
#9710
Conversation
isVoid
added
feature request
Codecov Report
@@ Coverage Diff @@
## branch-22.02 #9710 +/- ##
================================================
- Coverage 10.49% 10.39% -0.10%
================================================
Files 119 119
Lines 20305 20523 +218
================================================
+ Hits 2130 2134 +4
- Misses 18175 18389 +214
Continue to review full report at Codecov.
|
…pes in date_range
| return self.copy() | ||
|
|
||
| pd_offset = pd.tseries.frequencies.to_offset(offset) | ||
| to_search = op(pd.Timestamp(self._index._column[idx]), pd_offset) |
There was a problem hiding this comment.
What does the op callable do? and what is the operation being done in this line // what does to_search repr? jw
There was a problem hiding this comment.
If first is called, then op is add to compute the cut-off date counting from the first date in the column; if last, op is sub to compute the that from the last date. to_search is the cut-off date. Feel free to request changes if the naming/logic isn't very readable to your taste.
| if ( | ||
| idx == 0 | ||
| and not isinstance(pd_offset, pd.tseries.offsets.Tick) | ||
| and pd_offset.is_on_offset(pd.Timestamp(self._index[0])) |
There was a problem hiding this comment.
does is_on_offset check for datetimes values only at the end of the offset?
There was a problem hiding this comment.
I think pandas DateOffset can be MonthBegin or MonthEnd. Depending on the string to be either MS or M.
In [3]: pd.tseries.frequencies.to_offset('M')
Out[3]: <MonthEnd>
In [4]: pd.tseries.frequencies.to_offset('MS')
Out[4]: <MonthBegin>I believe is_on_offset checks if the given datetime falls on the given offset. There isn't much documentation about this but it seems like every kind of offset has this function.
Like Nano would always return true:
https://github.com/pandas-dev/pandas/blob/878a0225c648cb145949f78085a8ff3f902a1c20/pandas/_libs/tslibs/offsets.pyx#L834-L835
and BusinessDay would check if the weekday of the given date is monday to friday
https://github.com/pandas-dev/pandas/blob/878a0225c648cb145949f78085a8ff3f902a1c20/pandas/_libs/tslibs/offsets.pyx#L1445-L1448
and for MonthEnd I believe it resorts to the base method:
https://github.com/pandas-dev/pandas/blob/878a0225c648cb145949f78085a8ff3f902a1c20/pandas/_libs/tslibs/offsets.pyx#L650-L659
Note that + and - on the MonthEnd basically sets the date to the end of that month, same apply to other offsets.
karthikeyann
changed the title
first and last methodfirst and last method to IndexedFrame
|
@gpucibot merge |
closes #9600
This PR adds
firstandlastmethod toindexed_frame. This method only applies toIndexedFramewithDatetimeIndexand gathers the first or last rows within time range specified byoffsetargument.