Patterns for ZonedDateTime values

The ZonedDateTime type supports the following patterns:

Standard Patterns

The following standard patterns are supported:

  • G: General invariant pattern, based on ISO-8601 down to the second, but including the time zone ID and offset. This does not include the calendar ID, and can only be used for parsing when specified in conjunction with a time zone provider. It corresponds to the custom pattern of uuuu'-'MM'-'dd'T'HH':'mm':'ss z '('o<g>')'. This is the default format pattern.
  • F: Extended invariant pattern, based on ISO-8601 down to the nanosecond, but including the time zone ID and offset. This does not include the calendar ID, and can only be used for parsing when specified in conjunction with a time zone provider. It corresponds to the custom pattern of uuuu'-'MM'-'dd'T'HH':'mm':'ss;FFFFFFFFF z '('o<g>')'.

Custom Patterns

The custom format patterns for a zoned date and time are provided by combining the custom patterns for OffsetDateTime with the addition of two extra custom format specifiers: z and x.

z is used to parse or format that time zone identifier. When parsing, an IDateTimeZoneProvider is used to extract candidate identifiers and fetch time zones for them. The UTC+/-xx:xx format for fixed offset time zones is always valid, regardless of provider. The provider is part of the ZonedDateTimePattern, and a new pattern with a different provider can be created using the WithZoneProvider method. The provider is not used when formatting: the time zone identifier is simply used directly. Note that if you format a ZonedDateTime which uses a time zone from a different provider than the one in the pattern, you may not be able to parse it again with the same pattern. A null provider can be specified, in which case the pattern can only be used for formatting. (This is used in the ToString override, for example.)

x is used only for formatting; it includes the abbreviation associated with the time zone at the given time, such as "PST" or "PDT". This is format-only as abbreviations are often ambiguous; they are not a substitute for full time zone identifiers. Note that time zones from the TZDB provider only have abbreviations; time zones from the BCL provider have the standard/daylight names provided by TimeZoneInfo. For example, in the London time zone in summer, this will format as "BST" when using the TZDB-based zone, and "GMT Summer Time" when using the BCL-based zone.

When parsing, if the pattern does not contain the z specifier, the time zone from the default value is used. The standard patterns all use a default value with the UTC time zone.

If the pattern does not contain an offset specifier ("o<...>") the local date and time represented by the text is interpreted according to the ZoneLocalMappingResolver associated with the pattern. A new pattern can be created from an existing one, just with a different resolver, using the WithResolver method. If the resolver throws a SkippedTimeException or AmbiguousTimeException, these are converted into UnparsableValueException results. Note that a pattern without an offset specifier will always lead to potential data loss when used with time zones which aren't a single fixed offset, due to the normal issues of time zone transitions (typically for daylight saving time). A null resolver can be specified, in which case the pattern can only be used for formatting.

If the pattern does contain an offset specifier, then when parsing, the offset present in the text is validated against the time zone. By specifying both a time zone identifier and an offset, the ambiguity around time zone transitions is eliminated. Again, if the offset is invalid for the time zone at the given local date and time, an UnparsableValueException result is produced.

The ld<...>, lt<...> and l<...> specifiers from OffsetDateTime patterns are also supported for ZonedDateTime, to allow for standard date/time patterns to be used as part of the ZonedDateTime pattern.