Optimising changing column types - sql

Question

I've got an ETL script which changes the column types of a table to desired type and updates the data. The script works perfectly but I was wondering if there was a better & quicker way of doing it? Working in Redshift.
Current procedure;
CREATE or REPLACE PROCEDURE p_alter_staging_tbls() AS $$
DECLARE
row RECORD;
BEGIN
FOR row IN select * from
(
select distinct table_name, column_name,data_type from staging.staging_col_info_v a order by a.table_name asc
)
loop
EXECUTE 'ALTER TABLE staging.' || row.table_name || ' ' || 'ADD COLUMN ' || concat('new_',row.column_name) || ' ' || row.data_type ;
EXECUTE 'UPDATE staging.' || row.table_name || ' ' || 'SET ' || concat('new_',row.column_name) || ' ' || '=' || ' ' || row.column_name || '::' || row.data_type ;
EXECUTE 'ALTER TABLE staging.' || row.table_name || ' ' || 'DROP COLUMN ' || row.column_name ;
execute 'ALTER TABLE staging.' || row.table_name || ' ' || 'RENAME COLUMN '|| concat('new_',row.column_name) || ' ' || 'TO ' || row.column_name;
END LOOP;
RETURN;
END;
$$ LANGUAGE plpgsql;

Answer 1

For optimization, you can check if the data type needs to be changed or if you already have the correct type and size.
For this purpose you can use a query like this:
SELECT table_schema, table_name, column_name, data_type,
column_default, character_maximum_length, numeric_precision
FROM information_schema.columns
WHERE table_schema = 'staging'
AND table_name = row.table_name
AND column_name = row.column_name
and then check if "data_type" AND "numeric_precision" is different, in this case you can alter the column.
You can also consider to not create a new column, but you can use:
ALTER TABLE table_name
ALTER COLUMN column_name [SET DATA] TYPE new_data_type;
instead of:
EXECUTE 'ALTER TABLE staging.' || row.table_name || ' ' || 'ADD COLUMN ' || concat('new_',row.column_name) || ' ' || row.data_type ;
EXECUTE 'UPDATE staging.' || row.table_name || ' ' || 'SET ' || concat('new_',row.column_name) || ' ' || '=' || ' ' || row.column_name || '::' || row.data_type ;
EXECUTE 'ALTER TABLE staging.' || row.table_name || ' ' || 'DROP COLUMN ' || row.column_name ;
execute 'ALTER TABLE staging.' || row.table_name || ' ' || 'RENAME COLUMN '|| concat('new_',row.column_name) || ' ' || 'TO ' || row.column_name;
you can use this:
EXECUTE 'ALTER TABLE staging.' || row.table_name || ' ' || 'ALTER COLUMN '|| row.column_name || ' ' || 'TYPE ' || || row.data_type;