5
88%
master: 88%

Ran 17 Dec 2015 11:40AM UTC

Files 10

Run time 1s

Badge

Embed ▾

Committed 17 Dec 2015 11:27AM UTC coverage: 89.529% (-0.02%) from 89.547%

Job # 101.5

Build Type

push

travis-ci

Committed by

amacd31

Commit Message

Simplify date conversion code in irregular write

Attempt to improve performance of irregular series writes by simplifying
the date conversion and comparison code.

Timings of tests:

In []: date = pd.to_datetime(datetime(2015,1,1, 10, 15, 12, 10))

In []: %timeit -n 100000 -r 3 int(date.value / 1000000000)
100000 loops, best of 3: 432 ns per loop    # Selected for this commit

In []: %timeit -n 100000 -r 3 calendar.timegm(date.utctimetuple())
100000 loops, best of 3: 3.13 µs per loop   # Previous method

In []: %timeit -n 100000 -r 3 np.datetime64(date, 's').astype('int')
100000 loops, best of 3: 31.6 µs per loop   # Alternate method investigated

Signed-off-by: Andrew MacDonald <andrew@maccas.net>

Coverage Stats

513 of 573 relevant lines covered (89.53%)

0.9 hits per line

amacd31 / phildb / 101 / 5
88%
master: 88%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 101.5

amacd31 / phildb / 101 / 5 88% master: 88%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 101.5

amacd31 / phildb / 101 / 5
88%
master: 88%

README BADGES
x