通过 postgresql 中的循环添加多列

Question

I'm trying to create a new table in a schema that has multiple columns, named via an index.

CREATE TABLE rflux (pk SERIAL PRIMARY KEY NOT NULL);

Now I want to add new columns like col0 FLOAT, col1, col2, col3, .... up to colN.

I know I can do something like

ALTER TABLE rflux add column col0 FLOAT add column col1 FLOAT ... ;

but I don't want to type everything out, since I need to create ~4500 columns. I'm trying to implement this with loops but I can't quite get it working. Does anyone have any ideas? I've tried creating a function to do this ..

create function addColumns()
returns void
as $$
begin
for i in 1..10 loop
alter table rflux add column tmp float;
alter table rflux rename tmp to col||i;
end loop;
return;
end;
$$
language plpgsql;

then do select * from addColumns();

but I get errors when renaming the column to col||i , or even just when I try i. I'm not even sure if this is the best way to do this. How do I add multiple columns where I can increment the name of the column with a counter?

Edit..I know I can't do this with 4500 columns, but what's the solution to this problem if I wanted to do it for 10 columns, say?

Thanks.

Answer 1

your design would likely fit better into an array, hstore, or json data type. Adding 4500 columns is a nightmare waiting to happen.

Answer 2

If it can help:

-- VERSION : POSTGRESQL 9.3

-- FICTIVE TABLE #1

CREATE TABLE table_1 ("YEAR" int, "CODE_SP" text, "TOTAL" int);

INSERT INTO table_1 VALUES
(2035, 'TRUC', 2),
(2035, 'MUCH', 4),
(2036, 'FLIC', 7),
(2036, 'LORD', 2),
(2036, 'TRUC', 8),
(2037, 'MUCH', 2),
(2037, 'FLIC', 2),
(2037, 'FLAC', 5),
(2037, 'MORD', 9),
(2038, 'LOOP', 3),
(2038, 'MUCH', 3);

SELECT * FROM table_1;

-- FICTIVE TABLE #2

CREATE TABLE table_2 ("YEAR" int);

INSERT INTO table_2("YEAR")
SELECT serial
FROM generate_series(2035,2038,1) AS serial;

SELECT * FROM table_2;

-- LOOP FOR ADDING COLUMNS ON TABLE #2 

DO
$do$
    DECLARE colnames TEXT;
BEGIN
FOR colnames IN 
    SELECT "CODE_SP"
    FROM table_1
    GROUP BY "CODE_SP"
    ORDER BY "CODE_SP"
LOOP
    EXECUTE 'ALTER TABLE table_2 ADD COLUMN ' || quote_ident(colnames) || ' INT DEFAULT NULL;'; /* careful: in quoted text, the spaces are important */
END LOOP;
END
$do$;

-- LOOP FOR COMPLETTING CELLS ON TABLE #2 

DO
$do$
    DECLARE id_value TEXT;
BEGIN
FOR id_value IN
    SELECT "CODE_SP"
    FROM table_1
    GROUP BY "CODE_SP"
    ORDER BY "CODE_SP"
LOOP
    EXECUTE 'UPDATE table_2 SET ' || quote_ident(id_value) || ' = table_1."TOTAL" FROM table_1 WHERE table_1."CODE_SP" = ' || quote_literal(id_value) || ' AND table_1."YEAR" = table_2."YEAR";'; /* careful: in quoted text, the spaces are important */
END LOOP;
END
$do$;

Answer 3

One of the additional solutions that I did not find, but created myself.

do $$
declare 
    colnames text[];
    target_array text[]:= array[
                                ['table1','product_id'],
                                ['table2','product_id'],
                                ['table3','product_id']
                                ];  
    
begin
   FOREACH colnames SLICE 1 in  ARRAY target_array
   loop
        execute 'alter table '||colnames[1]||' 
                add  '||colnames[2]||' int
                DEFAULT NULL;';
        raise notice 'table: %, column: %', colnames[1], colnames[2];
   end loop;
end; $$;