How to make a function in postgresql with two parameters to return count from column table?

Question

I'm trying to do a function with two in parameters year and code from a table to return the count of the given code and year. The table is something like this:

CREATE TABLE orders (
    id INT NOT NULL, 
    code INT NOT NULL, 
    service_id INT NOT NULL, 
    status CHARACTER VARYING(50), 
    creation TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
    status_date TIMESTAMP,
    created_user INT
    );

I was able to return the count with the function below but when I give values to year and code in which there is no count it returns NULL but zero should be returned, how can I do that?:

CREATE OR REPLACE FUNCTION yearly_orders
    (year INT, order_cat orders.code%type) RETURNS INT AS $$
    
DECLARE
    total INT DEFAULT 0;
    
BEGIN
    SELECT COUNT(orders.code) INTO total
    FROM orders
    WHERE status = 'Requested' AND EXTRACT(YEAR FROM creation) = year::INT AND orders.code = order_cat
    GROUP BY year, orders.code;
    RETURN total;
END;
$$LANGUAGE plpgsql;

The data is something like:

INSERT INTO orders 
(id, code, service_id, status, creation, status_date, created_user) 
VALUES 
(100,2394,558151,'Requested','2019-06-16 11:12','2019-06-18 14:08',1);

Answer 1

This would work:

CREATE OR REPLACE FUNCTION yearly_orders(_year int, _order_cat int)
  RETURNS int
  LANGUAGE sql STABLE AS
$func$
SELECT count(*)::int
FROM   orders o
WHERE  o.status = 'Requested'
AND    o.creation >= to_timestamp(_year::text, 'YYYY')
AND    o.creation <  to_timestamp((_year + 1)::text, 'YYYY')
AND    o.code = _order_cat;
$func$;

The conversion from integer to timestamp really depends on what you pass exactly and what that's supposed to mean. You might want to_timestamp(_year::text, 'YY') instead ...

Consider the list in Akina's comment.
Start reading the manual about about SQL functions.
Have a look at PL/pgSQL functions after that.

Overview:

• Difference between language sql and language plpgsql in PostgreSQL functions

Answer 2

Seeing as the question has been reopened, I'll post my solution too.

'orders' Table

CREATE TABLE orders (
    id INT NOT NULL, 
    code INT NOT NULL, 
    service_id INT NOT NULL, 
    status CHARACTER VARYING(50), 
    creation TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
    status_date TIMESTAMP,
    created_user INT
    );

Records

INSERT INTO orders 
(id, code, service_id, status, creation, status_date, created_user) 
VALUES 
(100,2394,558151,'Requested','2019-06-16 11:12','2019-06-18 14:08',1),
(100,2394,558151,'Fulfilled','2018-06-16 11:12','2018-06-18 14:08',1),
(100,2394,558151,'Requested','2019-06-16 11:12','2019-06-18 14:08',1),
(100,2395,558152,'Requested','2019-06-16 11:12','2019-06-18 14:08',1),
(100,2396,558153,'Requested','2019-06-16 11:12','2019-06-18 14:08',1),
(100,2394,558151,'Requested','2019-06-16 11:12','2019-06-18 14:08',1);

Testing Your Statement

SELECT EXTRACT(YEAR FROM creation) AS year, 
                        orders.code, 
                        COUNT(orders.code) 
                FROM orders
                WHERE status = 'Requested' AND EXTRACT(YEAR FROM creation) = 2019 AND orders.code = 2394
                GROUP BY year, orders.code
                ORDER BY year ASC

Returns

year | code | count
:--- | ---: | ----:
2019 | 2394 |     3

Function

CREATE OR REPLACE FUNCTION 
yearly_orders(inyear numeric DEFAULT 0, incode INT DEFAULT 0) 
RETURNS TABLE (retyear numeric, retcode bigint, retcout bigint) AS 
$$
DECLARE
    retcount int DEFAULT 0;
    retcode int DEFAULT 0;
    retyear numeric DEFAULT 0;
BEGIN
    RETURN QUERY SELECT CAST(EXTRACT(year FROM creation) as numeric), 
                        CAST(orders.code as bigint), 
                        COUNT(orders.code)
                FROM orders
                WHERE status = 'Requested' 
                AND EXTRACT(YEAR FROM creation) = inyear 
                AND orders.code = incode
                GROUP BY EXTRACT(year FROM creation), orders.code
                ORDER BY retyear ASC;

END;
$$
LANGUAGE plpgsql;

Executing Function

select * from yearly_orders(2019, 2394);

Returns

retyear | retcode | retcout
------: | ------: | ------:
   2019 |    2394 |       3

db<>fiddle

The db<>fiddle to try this out can be found here.