I have a function like this
CREATE OR REPLACE FUNCTION factors_apart_sa()
RETURNS TABLE(surf_area integer, sa_factor numeric) AS
$BODY$
BEGIN
RETURN QUERY SELECT factors_apart_sa.surf_area, factors_apart_sa.sa_factor FROM factors_apart_sa;
END;
$BODY$
LANGUAGE plpgsql VOLATILE
COST 100
ROWS 1000;
After calling this function I get an error:
column reference "factors_apart_sa.surf_area" is ambiguous
How come is it still ambiguous if I have specified the table name.
Here is the table factors_apart_sa:
CREATE TABLE factors_apart_sa
(
factors_apar_sa_id bigserial NOT NULL,
surf_area integer,
sa_factor numeric(19,4) NOT NULL
)
Use a table alias:
CREATE OR REPLACE FUNCTION factors_apart_sa()
RETURNS TABLE(surf_area integer, sa_factor numeric) AS
$BODY$
BEGIN
RETURN QUERY SELECT f.surf_area, f.sa_factor FROM factors_apart_sa f;
END;
$BODY$
LANGUAGE plpgsql VOLATILE;
It is not a good idea to have the same names for a table and a function;
To call a function returnig set of rows place it in a FROM clause:
select f.*
from factors_apart_sa() f;
Related
I have created a function to delete multiple records.In our table contain id as type uuid.
We get the input is like array of ids.
CREATE OR REPLACE FUNCTION public.deletetVersion(item_list uuid[])
RETURNS TABLE(id uuid[])
LANGUAGE 'plpgsql'
COST 100
VOLATILE PARALLEL UNSAFE
ROWS 1000
AS $BODY$
BEGIN
RETURN QUERY
DELETE FROM version WHERE id = ANY(item_list);
END;
$BODY$;
SELECT * from deletetVersion(Array['b6ad1912-e4f1-4419-831a-c70df89ffd63','877898f0-2f3f-4890-a658-898e35ffee3a'])
But i got an error like:
Anyone please help me
ERROR: function deletetversion(text[]) does not exist
it is because the
Array['b6ad1912-e4f1-4419-831a-c70df89ffd63','877898f0-2f3f-4890-a658-898e35ffee3a']
is treated as text[]
try the following
Array['b6ad1912-e4f1-4419-831a-c70df89ffd63'::uuid,'877898f0-2f3f-4890-a658-898e35ffee3a'::uuid]
as a parameter to your function
for example
CREATE OR REPLACE FUNCTION public.test_uuid(item_list uuid[])
RETURNS TABLE(id uuid[])
LANGUAGE 'plpgsql'
COST 100
VOLATILE PARALLEL UNSAFE
ROWS 1000
AS $BODY$
BEGIN
RETURN QUERY
SELECT item_list;
END;
$BODY$;
SELECT * from test_uuid(Array['b6ad1912-e4f1-4419-831a-c70df89ffd63'::uuid])
In case of deletion
CREATE OR REPLACE FUNCTION public.test_uuid(item_list uuid[])
RETURNS VOID
LANGUAGE 'plpgsql'
COST 100
VOLATILE PARALLEL UNSAFE
ROWS 1000
AS $BODY$
BEGIN
RETURN QUERY
DELETE from tableName WHERE id = ANY(item_list);
END;
$BODY$;
Your function should return either setof uuid - i.e. a table of uuid-s - or uuid[]. I would prefer the first. You do not need PL/pgSQL, plain SQL is enough. So the function is:
create or replace function public.deletetVersion(item_list uuid[])
returns setof uuid language 'sql' as
$$
delete from version where id = any(item_list) returning id;
$$;
The version returning an array is a bit more complex:
create or replace function public.deletetVersion(item_list uuid[])
returns uuid[] language 'sql' as
$$
with t(d_id) as
(
delete from version where id = any(item_list) returning id
)
select array_agg(d_id) from t;
$$;
And - as #Ibrahimshamma says - you may need to cast the argument to uuid[].
My project used MS SQL, and postgresql so I can't change my parameter to another name, something like below
CREATE OR REPLACE FUNCTION myfunction( ReservationNo integer)
RETURNS TABLE("reservationno" integer) AS
$BODY$
begin
RETURN QUERY
select
rr.ReservationNo
FROM Mytable rr
WHERE rr.ReservationNo = myfunction.ReservationNo;
end
$BODY$
LANGUAGE plpgsql VOLATILE
COST 100
ROWS 1000;
ALTER FUNCTION myfunction(integer)
OWNER TO postgres;
But it give me an error
parameter name "reservationno" used more than once
How can I fix it?
You can change it to a SQL function, which doesn't have this problem:
CREATE OR REPLACE FUNCTION myfunction( ReservationNo integer)
RETURNS TABLE("reservationno" integer) AS
$BODY$
select rr.ReservationNo
FROM Mytable rr
WHERE rr.ReservationNo = myfunction.ReservationNo;
$BODY$
LANGUAGE sql stable
;
Suppose I have the following table, function and execution:
create table mytable (a INTEGER, b INTEGER);
create function test(q INTEGER)
returns table(a INTEGER, b INTEGER)
as
$body$
begin
return query select a,b from mytable;
end;
$body$
language plpgsql STABLE;
select * from test(1);
I get an 'ambiguous column name' error. I can get rid of it by changing the selection to "select t.a, t.b from mytable t" (per some similar-ish posts). But it seems very odd to have to qualify the column names when there is only 1 table in my query. I'm porting code that has quite a lot of stored procedures selecting from single tables (in various ways) and returning a table with columns that have the same name. Is there a better way than this of avoiding the error, and still having an output table with the same column names?
Thanks for any leads.
you can (you should) to use aliases.
create table mytable (a INTEGER, b INTEGER);
create function test(q INTEGER)
returns table(a INTEGER, b INTEGER)
as
$body$
begin
return query select mt.a, mt.b from mytable mt;
end;
$body$
language plpgsql STABLE;
select * from test(1);
maybe thats an option, using format and query execute:
( as the error says, it doesnt know which a to take, the pl/pgSQL variable or a columname )
create function test(q INTEGER)
returns table(a INTEGER, b INTEGER)
as
$body$
declare
lsQueryExecute text;
begin
lsQueryExecute = format('select a,b from mytable');
return query execute lsQueryExecute;
end;
$body$
language plpgsql STABLE;
I have a database that is named Dvdrental and in that database there is a function that is named public.film_not_in_stock, but I don't understand what it does.
-- Function: public.film_not_in_stock(integer, integer)
-- DROP FUNCTION public.film_not_in_stock(integer, integer);
CREATE OR REPLACE FUNCTION public.film_not_in_stock(
IN p_film_id integer,
IN p_store_id integer,
OUT p_film_count integer)
RETURNS SETOF integer AS
$BODY$
SELECT inventory_id
FROM inventory
WHERE film_id = $1
AND store_id = $2
AND NOT inventory_in_stock(inventory_id);
$BODY$
LANGUAGE sql VOLATILE
COST 100
ROWS 1000;
ALTER FUNCTION public.film_not_in_stock(integer, integer)
OWNER TO postgres;
It's a LANGUAGE SQL function. Not plpgsql.
SQL-language functions implicitly return the results of the last query they run.
Your function, if it was rewritten in plpgsql, would do
$BODY$
BEGIN
RETURN QUERY
SELECT inventory_id
FROM inventory
WHERE film_id = $1
AND store_id = $2
AND NOT inventory_in_stock(inventory_id);
$BODY$
LANGUAGE plpgsql VOLATILE
but in a LANGUAGE SQL function the RETURN is implicit.
I've got a postgresql stored procedure, which is returning an integer.
When I call that function, the result is returned with the function name as column name.
For example the name of the function is: "add-person". The column name, when invoking the function, is "add-person".
Is there a way to make the database return the integer with a self-choosen column name? For example "id"?
I think it is pretty easy, but I currently miss the forests for the trees..
Edit:
What i'd missed to tell, is that the return value is a variable, like so:
CREATE OR REPLACE FUNCTION "scheme"."add-person"(arggivenname character varying, argfamilyname character varying) RETURNS integer AS
$BODY$
DECLARE
varResponse integer;
BEGIN
-- Operations before
INSERT INTO "scheme"."table"
(
given_name,
family_name
)
VALUES
(
arggivenname,
argfamilyname
)
RETURNING
"id"
INTO
varResponse;
-- Operations after
RETURN varResponse;
END;
$BODY$
LANGUAGE plpgsql VOLATILE COST 100;
You can us the AS statement for that. That means:
Select add-person() AS yourcolumnname
To have a named column from a function it is necessary to create a type and return that type from the function
create type mytype as (mycolumn integer);
create or replace function ri()
returns mytype as $$
select 1;
$$ language sql;
select * from ri();
mycolumn
----------
1
Edit
Or much simpler without the type creation as in #pozs comment:
create or replace function ri(out mycolumn integer)
as $$
select 1;
$$ language sql;