Why iam getting this error while executing Postgres function - postgresql

Question

CREATE OR REPLACE FUNCTION public.flowrate7(
)
RETURNS TABLE(oms_id integer, flowrate numeric, chakno character varying)
LANGUAGE 'plpgsql'
COST 100
VOLATILE
ROWS 1000
AS $BODY$
declare temp_omsId integer;
declare temp_flowrate numeric(20,3);
declare temp_chakno varchar(100);
begin
DROP TABLE IF EXISTS tbl_oms;
DROP TABLE IF EXISTS tbl_calFlow;
CREATE temporary TABLE tbl_oms(omsid__ integer) ON COMMIT DELETE ROWS;
CREATE temporary TABLE tbl_calFlow(OmsId_ integer,FlowRate_ numeric(20,3),ChakNo_ varchar(100)) ON COMMIT DELETE ROWS;
insert into tbl_oms (select OmsId from MstOms);
while (select count(*) from tbl_oms) <> 0 LOOP
select temp_omsId = omsid__ from tbl_oms LIMIT 1;
temp_flowrate = (select (case when(InLetPressure > 0.5) then 1 else 0 end) from MstOms where OmsId = temp_omsId);
temp_chakno = (select ChakNo from MstOms where OmsId = temp_omsId);
insert into tbl_calFlow values (temp_omsId,temp_flowrate,temp_chakno);
delete from tbl_oms where omsid__ = temp_omsId;
END LOOP;
return query (select OmsId_,FlowRate_,ChakNo_ from tbl_calFlow);
end;
$BODY$;
ERROR: query has no destination for result data
HINT: If you want to discard the results of a SELECT, use PERFORM instead.
CONTEXT: PL/pgSQL function flowrate7() line 19 at SQL statement
SQL state: 42601

Answer 1

You are retrieving values into the variables incorrectly. To store the value of a query result into a variable (or two) you need to use select .. into variable ...
CREATE OR REPLACE FUNCTION public.flowrate7()
RETURNS TABLE(oms_id integer, flowrate numeric, chakno character varying)
LANGUAGE plpgsql
AS $BODY$
declare
temp_omsId integer;
temp_flowrate numeric(20,3);
temp_chakno varchar(100);
begin
DROP TABLE IF EXISTS tbl_oms;
DROP TABLE IF EXISTS tbl_calFlow;
CREATE temporary TABLE tbl_oms(omsid__ integer) ON COMMIT DELETE ROWS;
CREATE temporary TABLE tbl_calFlow(OmsId_ integer,FlowRate_ numeric(20,3),ChakNo_ varchar(100)) ON COMMIT DELETE ROWS;
insert into tbl_oms
select OmsId
from MstOms;
while (select count(*) from tbl_oms) <> 0 LOOP
select omsid__
into temp_omsId --<< here
from tbl_oms LIMIT 1;
select case when inletpressure> 0.5 then 1 else 0 end, chakno
into temp_flowrate, temp_chakno --<< here
from MstOms
where omsid = temp_omsId;
insert into tbl_calFlow values (temp_omsId,temp_flowrate,temp_chakno);
delete from tbl_oms where omsid__ = temp_omsId;
END LOOP;
return query select omsid_, flowrate_, chakno_
from tbl_calflow;
end;
$BODY$;
However the processing of that function is unnecessarily complicated
if first copies all rows MstOms to tmp_MstOms
retrieve the ID for one row from tbl_oms
retrieve one row from MstOms calculating the flowrate
stores that one row in the temp table
deletes the just processed row from the (other) temp table
counts the number or rows in tbl_oms and if that is not zero moves on to the "next" row
This is an extremely inefficient and complicate way of calculating a simple value which will not scale well for large tables.
Doing a row-by-row processing of tables in the database is an anti-pattern to begin with (and doing that by deleting, inserting and counting rows is making that even slower).
This is not how things are done in SQL. The whole inefficient loop can be replaced with a single query which also allows you to change the whole thing to a SQL function.
CREATE OR REPLACE FUNCTION public.flowrate7()
RETURNS TABLE(oms_id integer, flowrate numeric, chakno character varying)
LANGUAGE sql
AS $BODY$
select omsid,
case when inletpressure> 0.5 then 1 else 0 end as flowrate,
chakno
from mstoms;
$BODY$;
Actually a view would be more appropriate for this.
create or replace view flowrate7
as
select omsid,
case when inletpressure> 0.5 then 1 else 0 end as flowrate,
chakno
from mstoms;