Optimizing pgSQL function performance - postgresql

Question

I really need advice. I'm trying to use the function : pivot_subrecords_set_table_1
In the following query:
SELECT rec.id AS key,
rec.primaryurn AS urn,
rec.mx_syscreatedby AS creared_by,
(SELECT mf_mxdate_str_to_date (epi.mx_epizoddate)) AS date_of_creation,
epi.mx_epizodenotes AS notes,
epi.mx_effectscountry AS country,
epi.mx_effectsregion AS region,
epi.mx_effectscity AS city,
(SELECT get_epizod_subrecords (epi.id)) AS epizod_type
(pivot_subrecords_set_table_1 (epi.id)).idx,
(pivot_subrecords_set_table_1 (epi.id)).upn_1 AS upn_1,
(pivot_subrecords_set_table_1 (epi.id)).status_1
FROM mxt_recordheader AS rec, mt_epizod AS epi, mt_epizod_subrecord AS eps
WHERE rec.entitytype = 'epizod'
AND rec.logicalserverprefix = 'EA'
AND epi.id = rec.id;
The function looks like this:
CREATE OR REPLACE FUNCTION public.pivot_subrecords_set_table_1(id bigint,
OUT idx bigint, OUT upn_1 text, OUT status_1 text, OUT upn_2 text, OUT
status_2 text, OUT upn_3 text, OUT status_3 text, OUT upn_4 text, OUT
status_4 text, OUT upn_5 text, OUT status_5 text, OUT upn_6 text, OUT
status_6 text, OUT upn_7 text, OUT status_7 text, OUT upn_8 text, OUT
status_8 text, OUT upn_9 text, OUT status_9 text, OUT upn_10 text, OUT
status_10 text)
RETURNS SETOF record
LANGUAGE plpgsql
AS $function$
DECLARE
key ALIAS FOR $1;
BEGIN
RETURN QUERY
WITH sub_upn AS (SELECT * FROM crosstab
('SELECT id, mx_osobastatus, mx_osobaupn FROM mt_epizod_subrecord
WHERE mt_epizod_subrecord.mx_syscategory = ''tableepizodinfo''
AND mt_epizod_subrecord.id = '|| key ||' GROUP BY id, mx_osobaupn,
mx_osobastatus
ORDER BY id') AS ct
(id bigint, UPN_1 text, UPN_2 text, UPN_3 text, UPN_4 text, UPN_5 text, UPN_6
text, UPN_7 text, UPN_8 text ,UPN_9 text, UPN_10 text)
),
sub_stat AS
(SELECT * FROM crosstab
('SELECT id, mx_osobaupn, mx_osobastatus FROM mt_epizod_subrecord
WHERE mt_epizod_subrecord.mx_syscategory = ''tableepizodinfo''
AND mt_epizod_subrecord.id = '|| key ||' GROUP BY 1,2,3 ORDER BY id ')
AS cn (id bigint, status_1 text, status_2 text, status_3 text, status_4 text,
status_5 text, status_6 text, status_7 text, status_8 text,
status_9 text, status_10 text)
)
SELECT
supn.id,
supn.UPN_1,
sust.status_1,
supn.UPN_2,
sust.status_2,
supn.UPN_3,
sust.status_3,
supn.UPN_4,
sust.status_4,
supn.UPN_5,
sust.status_5,
supn.UPN_6,
sust.status_6,
supn.UPN_7,
sust.status_7,
supn.UPN_8,
sust.status_8,
supn.UPN_9,
sust.status_9,
supn.UPN_10,
sust.status_10
FROM sub_upn supn
JOIN sub_stat sust ON supn.id = sust.id;
END;
$function$
But the performance of the function is very low.
How can I increase the performance of a function or can I rewrite the query in which it is used?

Answer 1

Use plprofiler to figure out what is slow in your function and optimize that.