Обсуждение: rolling window without aggregation
Hi SQL experts,
I’ve got a question here, is that possible to implement a window function without aggregation? Any SQL could get below desired result?
For example:
Table input
date | id
------------+--------
2014-04-26 | A
2014-05-03 | B
2014-05-10 | C
2014-05-17 | D
2014-05-24 | E
2014-05-31 | F
Expected output, use 2 week roll up as an example:
date | id
------------+--------
2014-04-26 | A
2014-05-03 | A
2014-05-03 | B
2014-05-10 | B
2014-05-10 | C
2014-05-17 | C
2014-05-17 | D
2014-05-24 | D
2014-05-24 | E
2014-05-31 | E
2014-05-31 | F
Thanks,
Suya
Huang, Suya wrote > Hi SQL experts, > > I've got a question here, is that possible to implement a window function > without aggregation? Any SQL could get below desired result? > > For example: > > Table input > date | id > ------------+-------- > 2014-04-26 | A > 2014-05-03 | B > 2014-05-10 | C > 2014-05-17 | D > 2014-05-24 | E > 2014-05-31 | F > > Expected output, use 2 week roll up as an example: > date | id > ------------+-------- > 2014-04-26 | A > 2014-05-03 | A > 2014-05-03 | B > 2014-05-10 | B > 2014-05-10 | C > 2014-05-17 | C > 2014-05-17 | D > 2014-05-24 | D > 2014-05-24 | E > 2014-05-31 | E > 2014-05-31 | F > > > > Thanks, > Suya Use the lead() function to create a second column. Then write a UNION ALL query to covert the two columns into one. David J. -- View this message in context: http://postgresql.nabble.com/rolling-window-without-aggregation-tp5829344p5829345.html Sent from the PostgreSQL - sql mailing list archive at Nabble.com.