Auto Date Parser

Detect the format of a string date column automatically and convert it to a real Date or Datetime — no need to know or specify the format up front.

Detection is done per row: the node tries a battery of ~50 common date/time formats and keeps the first that matches each value. Because the match is per-value, a single column may even contain a mix of formats.

Inputs

Port	Description
input[0]	A table containing a string column of dates.

Settings

Setting	Type	Default	Description
Date column	column (String)	—	The string column to parse. (required)
Convert to	choice	Date	Output as a Date (day precision) or Datetime (with time).
Output column	text	(empty)	Name for the parsed column. Leave empty to overwrite the source column.
Day first	toggle	off	For ambiguous values like 03/04/2021, read as day/month (4 March) instead of month/day.
On unparseable value	choice	Set to null	Set to null leaves bad values empty; Raise an error fails the node and reports up to 5 offending values.

Output

The chosen column, converted to Date or Datetime. Values that no format matches become null (or raise, depending on On unparseable value).

How detection works

Formats are tried in priority order and coalesced (first match wins per row):

1. Unambiguous / ISO first — 2021-03-04, 2021-03-04T10:30:00, 2021-03-04T10:30:00+02:00, 2021/03/04, 20210304, …
2. Month names — March 4, 2021, 4 Mar 2021, 04-Mar-2021, …
3. Ambiguous numeric — 04/03/2021, 03-04-21, 03.04.2021 across /, -, . separators and 2-/4-digit years, with or without a time part. The Day first toggle decides the order these are tried in.

Timezone-aware values (e.g. +02:00) are normalized to UTC and returned as naive datetimes, so the whole column shares one dtype.

Examples

With Convert to = Date, Day first = off:

raw_date	→ parsed_date
2021-03-04	2021-03-04
04/03/2021	2021-04-03 (month-first)
March 4, 2021	2021-03-04
2021-03-04T10:30:00	2021-03-04
20210304	2021-03-04
not a date	null

Turning Day first = on flips the ambiguous case: 04/03/2021 → 2021-03-04 (4 March).

Notes & edge cases

• Mixed formats in one column are supported — each value is matched independently.
• Ambiguous dates (e.g. 03/04/2021) are genuinely undecidable from the value alone; use Day first to pick the interpretation.
• 2-digit years with dashes: 04-03-21 parses to year 0004 (the 4-digit year pattern claims the leading 04 before the 2-digit variant is tried). If your data uses 2-digit years, prefer an unambiguous 4-digit source.
• Runs locally (in the worker) — no kernel or Docker required.

Metadata

• Category: Date & Time
• Author: Edwardvaneechoud
• Version: 1.0.0
• Tags: date, datetime, parsing, cleaning