Thanks for replying.... I know it's complicated to follow it (even if formatted ) , but probably I did not emphasize enough that by using unnest(ids) in the second query, the distinct count for company is not computed correctly anymore. Focusing on the used select clauses is important, they are just distinct counts, it would have been logical to stay the same for both queries, ignoring the unnest(ids)(the data set is in the milions of rows so you need to take my word that the first query return the correct values :) ).
Anyway, will try to test it with latest version, also will try to provide some sql test case for this, but again... There surely is an issue with the unnesting arrays (at least in the relase that I'm using)
Seb
On Fri, May 12, 2017 at 5:20 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
sebastian.calbaza@hgdata.com writes: > Below is a query that we are using to calculate some counts: > * first version of the query lacks ```unnest(ids) as id,```, but second > one has it > * ```companies``` count value is incorrect for the second one, first > query has the correct value
My first suggestion would be to see if updating to 9.6.3 fixes it. If not, please try to create a self-contained test case. These queries are unreadable, and without the underlying data, nobody else can even tell whether the answers are wrong or not.
