Skip to content

parse fails with subsecond decimals with more than 9 digits #935

@dr-eme

Description

@dr-eme
  • I am on the latest Pendulum version.

  • I have searched the issues of this repo and believe that this is not a duplicate.

  • OS version and name: Win10

  • Pendulum version: 3.1.0

Issue

pendulum.parse() trhows ParseError when parsing an ISO8601 formatted date with subsecond decimals with more than 9 digits, independent on whether time offset is present or not.

For instance, the following parse attempt work (truncating decimals to microsecond resolution)

>>> parse("2001-01-01T12:34:56.123456789")
DateTime(2001, 1, 1, 12, 34, 56, 123456, tzinfo=Timezone('UTC'))
>>> parse("2001-01-01T12:34:56.123456789Z")
DateTime(2001, 1, 1, 12, 34, 56, 123456, tzinfo=Timezone('UTC'))
>>> parse("2001-01-01T12:34:56.123456789+00:00")
DateTime(2001, 1, 1, 12, 34, 56, 123456, tzinfo=FixedTimezone(0, name="+00:00"))
>>> parse("2001-01-01T12:34:56.123456789+0000")
DateTime(2001, 1, 1, 12, 34, 56, 123456, tzinfo=FixedTimezone(0, name="+00:00"))

but the same with an extra digit (even if it's 0) fail

parse("2001-01-01T12:34:56.1234567890")
Traceback (most recent call last):
  File "<python-input-1>", line 1, in <module>
    parse("2001-01-01T12:34:56.1234567890")
    ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".venv\Lib\site-packages\pendulum\parser.py", line 36, in parse
    return _parse(text, **options)
  File ".venv\Lib\site-packages\pendulum\parser.py", line 51, in _parse
    parsed = base_parse(text, **options)
  File ".venv\Lib\site-packages\pendulum\parsing\__init__.py", line 78, in parse
    return _normalize(_parse(text, **_options), **_options)
                      ~~~~~~^^^^^^^^^^^^^^^^^^
  File ".venv\Lib\site-packages\pendulum\parsing\__init__.py", line 125, in _parse
    raise ParserError(f"Unable to parse string [{text}]")
pendulum.parsing.exceptions.ParserError: Unable to parse string [2001-01-01T12:34:56.1234567890]
```

It is expected that truncation (or even better rounding) to microseconds is applied independent of the number of digits present in the string.

The issue also happens when running in Linux (e.g., using the latest python3.14-slim container).

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions