How does the Postgres specification define the requirements for the columns other than the on() column ? Is it "one row at (pseudo-)random from all the rows which match this specific value of the on() column", or is there any indication that the other columns could ALL be chosen randomly so that the output wasn't a row that actually existed, or is there any indication that the result will be deterministic in a particular way ? Regards Jonathan Lewis http://jonathanlewis.wordpress.com/all_postings Author: Oracle Core (Apress 2011) http://www.apress.com/9781430239543 ----- Original Message ----- From: "jo" <jose.soares@xxxxxxxxxxxxxx> To: "ORACLE-L" <oracle-l@xxxxxxxxxxxxx> Sent: Thursday, October 18, 2012 7:38 AM Subject: distinct on |I found an interesting non standard feature in postgres to distinguish | one column in a query. | How can I do this in Oracle? | | | select distinct on(a) * from test; | a | b | ----+----- | a | bb | aa | bbb | ab | bbb | (3 rows) | -- //www.freelists.org/webpage/oracle-l