Re: query has no destination for result data

On 3/5/19 11:45 PM, Rob Sargent wrote:
> I’m using 10.7. Does an empty result set generate this error by any chance.
>
> One of my plpgsql functions is now throwing this error:
>
> select * from genome_threshold_mono('11-O3C.pbs','1-O3C_chr',1.96,
> 1000000);

The above is the function you started with.
See below for more.

> NOTICE:  group id is 5eed8d65-d39a-4f72-97a3-ca391b84880d
> NOTICE:  New threshold: 66128154-d128-4e66-bb8e-e9c9ee5ae89d
> NOTICE:  doing chrom 11
> NOTICE:  2019-03-06 00:21:17.253375-07: markerset id is
> 9a8f7487-bd64-4d43-9adf-5ae1c6744e60(1-O3C_chr11.Loci.ld), people
> (5eed8d65-d39a-4f72-97a3-ca391b84880d) id is 11-O3C.pbs
> NOTICE:  table "collected" does not exist, skipping
> NOTICE:  table "mrkidx" does not exist, skipping
> NOTICE:  2019-03-06 00:21:17.295142-07: working with 28607 markers
> NOTICE:  2019-03-06 00:21:17.383835-07: added 3514 segments to
> imputed_pvalue_t
> ERROR:  query has no destination for result data
> HINT:  If you want to discard the results of a SELECT, use PERFORM
> instead.

Below you have two other functions in play:

optimal_pvalue_mono()
genome_threshold_mono()

If I am following correctly it is line 30 in genome_threshold_mono()
you want to take a look at.

> CONTEXT:  PL/pgSQL function
> optimal_pvalue_mono(text,text,integer,double precision,integer) line
> 65 at SQL statement
> SQL statement "insert into goptset
>         select * from optimal_pvalue_mono(people_name, mvec.name,
> mvec.chrom, conf, maxi)"
> PL/pgSQL function genome_pvalue_mono(text,text,double
> precision,integer) line 19 at SQL statement
> SQL statement "insert into threshold_segment(id,threshold_id,
> segment_id, smooth_pvalue)
>          select uuid_generate_v4(), tid, f.segment_id, f.pval
>          from genome_pvalue_mono(pbs_name, markers_rx, conf, maxi)
> as f"
> PL/pgSQL function genome_threshold_mono(text,text,double
> precision,integer) line 30 at SQL statement
>
>
> The code referenced at line 65 is the last line in a "for row in query”
> construct as follows:
>
>     54 for segp in
>     55   select s.id <http://s.id>, s.firstmarker, s.lastmarker,
>     56          v.ipv,
>     57          array_length(p.probands,1) as pbs,
>     58          s.lastmarker - s.firstmarker as mks
>     59   from segment s
>     60        join imputed_pvalue_t v on s.id <http://s.id> =
> v.segment_id
>     61        join probandset p on s.probandset_id = p.id <http://p.id>
>     62        join probandset_group_member m on p.id <http://p.id>
> = m.member_id
>     63   where s.markerset_id = mkset
>     64         and m.group_id = pbsgid
>     65   order by ipv, pbs, mks
>     66 LOOP
>
> Plugging in the appropriate values for an example run generates a proper
> dataset (~1300 rows)as far as I can tell.
> This construct had been working until recent changes but I cannot relate
> the message to any deformity in the current schema or code.
> Any pointers appreciated.
>

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com