CLN: to_datetime internals

[mroeschke]

• tests passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff

The internals of to_datetime is getting a bit unwieldy, so split out the _convert_listlike logic and origin shifting logic to _convert_listlike_datetime and adjust_to_origin methods respectively outside of to_datetime. The logic was not changed.

[pep8speaks]

Hello @mroeschke! Thanks for updating the PR.

• In the file pandas/core/tools/datetimes.py, following are the PEP8 issues :

Line 243:17: E722 do not use bare except'
Line 315:9: E722 do not use bare except'

Comment last updated on July 03, 2018 at 05:24 Hours UTC

[codecov]

Codecov Report

Merging #21702 into master will increase coverage by <.01%.
The diff coverage is 88.88%.

@@            Coverage Diff             @@
##           master   #21702      +/-   ##
==========================================
+ Coverage    91.9%   91.91%   +<.01%     
==========================================
  Files         158      158              
  Lines       49690    49695       +5     
==========================================
+ Hits        45670    45675       +5     
  Misses       4020     4020

Flag	Coverage Δ
#multiple	90.28% <88.88%> (ø)	⬆️
#single	41.95% <51.28%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/tools/datetimes.py	85.22% <88.88%> (+0.23%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7cd2679...e9320f1. Read the comment docs.

[jreback]

lgtm. can you do a perf check to make sure nothing changed (as the caching logic was lightly touched here)

[jorisvandenbossche]

The tz here was not used?

[mroeschke]

tz was passed into the convert_listlike function further down, but now I am embedding it into convert_listlike with functools.partial in to_datetime

[jbrockmendel]

+1 for separating this out

[jbrockmendel]

Is it necessarily an ndarray? Couldn't it be a Timestamp?

need newline before Returns

[mroeschke]

Good catch, yeah this can be a scalar value. Will fix that tonight.

[mroeschke]

My asv setup is still a little broken, but here's a benchmark showing no performance hit (with cache)

In [1]: from pandas import *

In [3]: N = 100

In [4]: dup_string_with_tz = ['2000-02-11 15:00:00-0800'] * N

# this branch
In [5]: %timeit to_datetime(dup_string_with_tz, cache=True)
1.31 ms ± 44.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

In [6]: %timeit to_datetime(dup_string_with_tz, cache=False)
2.3 ms ± 41.2 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

# Master
In [4]: %timeit to_datetime(dup_string_with_tz, cache=True)
1.29 ms ± 13.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

In [5]: %timeit to_datetime(dup_string_with_tz, cache=False)
2.31 ms ± 68.9 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

[jreback]

thanks @mroeschke happily take refactorings to clean up things / make more readable