Big Query:
I'm looking to assign a 'unique combination id' if a person and date are an exact match. So for each specific time a 'person_id' and a specific datetime are the same, they are assigned a number overall (time will always be 00:00:00 but just has to be there for records). This will be a unique number overall, not for a count for each individual. Hope that makes sense.
Desired result below. So you'll see row 1 and 5 are the same unique combination, so have the same unique_combination_id.
person_id | Datetime | unique combination_id |
---|---|---|
1234 | 2016-04-01T00:00:00 | 1 |
1234 | 2016-05-04T00:00:00 | 2 |
9102 | 2018-05-17T00:00:00 | 3 |
5678 | 2019-09-01T00:00:00 | 4 |
1234 | 2016-04-01T00:00:00 | 1 |
CodePudding user response:
Using DENSE_RANK()
should do the trick:
SELECT *,
DENSE_RANK() OVER(ORDER by person_id, datetime) as unique_combination_id
FROM tbl;