nr.parsing.date

A fast, regular-expression based library for parsing dates, plus support for ISO 8601 durations.

Requirements

• Python 3.6+

Supported Date & Time Formats

• %Y – 4 digit year
• %m – 2 digit month
• %d – 2 digit day
• %H – 2 digit hour
• %M – 2 digit minute
• %S – 2 digit second
• %f – arbitrary precision milliseconds
• %z – timezone offset ([+-]\d\d:?\d\d offset or Z for UTC)

Built-in format collections

• ISO_8601 (see ISO 8601 on Wikipedia)
• JAVA_OFFSET_DATETIME (see OffsetDateTime class on the Java 8 API documentation)

Features

• Easily extensible to support more date/time format options
• Date/time formats can use an extended regex-style mode to mark format options as optional (e.g. the two formats %Y and %Y-%m can be expressed in a single regex-style format string as %Y(-%m)?)

Quickstart

from nr.parsing.date import duration, ISO_8601
ISO_8601.parse('2021-04-21T10:13:00.124+0000')
duration.parse('P3Y6M4DT12H30M5S')

Benchmark

The below benchmark compares the performance of testing various format-strings for ISO-8601 dates using the standard library, dateutil.parser.parse(), dateutil.parser.isoparse() and the nr.parsing.date.ISO_8601.parse_datetime() function.

Conclusion: Faster than the standard library but with the same flexibility (except for the missing support for most uncommon format options).

asv run
· Creating environments
· Discovering benchmarks
· Running 5 total benchmarks (1 commits * 1 environments * 5 benchmarks)
[  0.00%] · For nr.parsing.date commit dd35e795 <develop>:
[  0.00%] ·· Benchmarking virtualenv-py3.8-pandas-python-dateutil
[ 10.00%] ··· Running (benchmarks.DatetimeParsingSuite.time_datetime_datetime_strptime--).....
[ 60.00%] ··· benchmarks.DatetimeParsingSuite.time_datetime_datetime_strptime                                     2.22±0.3ms
[ 70.00%] ··· benchmarks.DatetimeParsingSuite.time_datetime_datetime_strptime_reversed                           2.12±0.08ms
[ 80.00%] ··· benchmarks.DatetimeParsingSuite.time_dateutil_parser_isoparse                                      1.46±0.02ms
[ 90.00%] ··· benchmarks.DatetimeParsingSuite.time_dateutil_parser_parse                                          2.77±0.1ms
[100.00%] ··· benchmarks.DatetimeParsingSuite.time_nr_parsing_date_ISO_8601_parse_datetime                       1.62±0.03ms

Copyright © 2020 Niklas Rosenstein