Timestamps and Datetime strings in Faxe
In Faxe we deal a lot with timestamps and datetime strings, in fact, every message that flows through all the nodes, has a timestamp with it.
Its inner representation of points in time is always an integer denoting the milliseconds that past since 1.1.1970. There is currently no timezone handling in Faxe.
Datetime string parsing
Faxe uses its own library to parse datetime strings (from the outside) into it's internal timestamp format. This is done in different nodes such as the mqtt_subscribe and the amqp_consume node. Furthermore, datetime parsing can be done in lambda-expression with the dt_parse function.
When parsing datetime strings with subsecond values smaller than milliseconds (microseconds and nanoseconds), if present, will be rounded to millisecond values.
Parsing examples
| datetime | matching format string |
|---|---|
| 'Mon, 15 Jun 2009 20:45:30 GMT' | 'a, d b Y H:M:S Z' |
| '19.08.01 17:33:44,867000 ' | 'd.m.y H:M:S,u ' |
| '8/28/2033 8:03:45.576000 PM' | 'n/d/Y l:M:S.u p' |
Conversion Specification
The format used to parse and format dates closely resembles the one used
in strftime()[1]. The most notable exception is that meaningful characters
are not prefixed with a percentage sign (%) in datestring.
Characters not matching a conversion specification will be copied to the output verbatim when formatting and matched against input when parsing. Meaningful characters can be escaped with a backslash (\).
| Character Sequence | Parsing | Formatting | Description |
|---|---|---|---|
| a | Yes* | Yes | Abbreviated weekday name ("Mon", "Tue") |
| A | Yes | Yes | Weekday name ("Monday", "Tuesday") |
| b | Yes* | Yes | Abbreviated month name ("Jan", "Feb") |
| B | Yes* | Yes | Month name ("January", "February") |
| d | Yes | Yes | Day of month with leading zero ("01", "31") |
| e | Yes | Yes | Day of month without leading zero ("1", "31") |
| F | No | Yes | ISO 8601 date format (shortcut for "Y-m-d") |
| H | Yes | Yes | Hour (24 hours) with leading zero ("01", "23") |
| I | Yes | Yes | Hour (12 hours) with leading zero ("01", "11") |
| k | Yes | Yes | Hour (24 hours) without leading zero ("1", "23") |
| l | Yes | Yes | Hour (12 hours) without leading zero ("1", "11") |
| m | Yes | Yes | Month with leading zero ("1", "12") |
| M | Yes | Yes | Minute with leading zero ("00", "59") |
| n | Yes | Yes | Month without leading zero ("1", "12") |
| o | Yes | Yes | Ordinal number suffix abbreviation (st, nd, rd, th) |
| p | Yes* | Yes | AM/PM |
| P | Yes* | Yes | a.m./p.m. |
| R | No | Yes | The time as H:M (24 hour format) ("23:59") |
| S | Yes | Yes | Seconds with leading zero ("00", "59") |
| T | No | Yes | The time as H:M:S (24 hour format) ("23:49:49") |
| c | Yes | No | Milliseconds, 3 digits with leading zero ("034") |
| u | Yes | Yes | Microseconds, 6 digits with leading zero ("000034") |
| f | Yes | No | Nanoseconds, 9 digits with leading zero ("000034000") |
| y | Yes** | Yes | Year without century ("02", "12") |
| Y | Yes | Yes | Year including century ("2002", "2012") |
| z | Yes | No | UTC offset (+0100, +01:00, +01, -0100, -01:00, -01) |
| Z | Yes | No | Abbreviated timezone (UTC, GMT, CET etc) |
* Case-insensitive when parsing
** Falls back on current century of system when parsing years without century.
Special formats
| dt_format | description | example |
|---|---|---|
| 'millisecond' | timestamp UTC in milliseconds | 1565343079000 |
| 'microsecond' | timestamp UTC in microseconds | 1565343079000123 |
| 'nanosecond' | timestamp UTC in nanoseconds | 1565343079000123456 |
| 'second' | timestamp UTC in seconds | 1565343079 |
| 'float_nano' | timestamp UTC float with nanosecond precision | 1565343079.173588126 |
| 'float_micro' | timestamp UTC float with microsecond precision | 1565343079.173588 |
| 'float_millisecond' | timestamp UTC float with millisecond precision | 1565343079.173 |
| 'ISO8601' | ISO8601 Datetime format string | '2011-10-05T14:48:00.000Z' |
| 'RFC3339' | RFC3339 Datetime format string | '2018-02-01 15:18:02.088Z' |