PostgreSQL function returns one column instead of 4 columns [duplicate]

Question

I have a PostgreSQL function that returns one column instead of 4 columns like it should. Moreover, all four data points are nested within one column and the updates were not performed. Can you please review my code and offer suggestions?

code:

-- FUNCTION: public.update_funcnull()

-- DROP FUNCTION IF EXISTS public.update_funcnull();

CREATE OR REPLACE FUNCTION public.update_funcnull(
    )
    RETURNS TABLE(ride_id text, duration interval, lat_change real, long_change real) 
    LANGUAGE 'plpgsql'
    COST 100
    VOLATILE PARALLEL UNSAFE
    ROWS 1000

AS $BODY$
BEGIN
    UPDATE bikeshare3 SET duration = NULL;
    UPDATE bikeshare3 SET lat_change = NULL;
    UPDATE bikeshare3 SET long_change = NULL;
    RETURN QUERY SELECT ride_id, duration, lat_change, long_change FROM bikeshare3 LIMIT 5;
END;
$BODY$;

ALTER FUNCTION public.update_funcnull()
    OWNER TO postgres;

Answer 1

You should to call function with "tabular context":

SELECT * FROM public.update_funcnull();

Tabular functions (or Set Returning functions in PostgreSQL terminology) returns composite values in "scalar context".

CREATE OR REPLACE FUNCTION public.foo()
 RETURNS TABLE(a integer, b integer)
 LANGUAGE plpgsql
AS $function$
BEGIN
  a := 1; b := 2;
  RETURN NEXT;
  RETURN NEXT;
END;
$function$;

(2022-04-18 05:55:32) postgres=# SELECT foo();
?????????
?  foo  ?
?????????
? (1,2) ?
? (1,2) ?
?????????
(2 rows)

(2022-04-18 05:55:37) postgres=# SELECT * FROM foo();
?????????
? a ? b ?
?????????
? 1 ? 2 ?
? 1 ? 2 ?
?????????
(2 rows)

Composite value can be unpacked, but packing to composite and unpacking has some performance impact:

(2022-04-18 05:55:43) postgres=# SELECT (foo()).*;
?????????
? a ? b ?
?????????
? 1 ? 2 ?
? 1 ? 2 ?
?????????
(2 rows)