SQL Query loop in Postgresql - sql

Question

I have DB in Postgresql 12 and PgAdmin 4 and a lot of queries like:
update public.register set address = replace(address,'Á','A') where address like '%Á%';
update public.register set address = replace(address,'Â','A') where address like '%Â%';
update public.register set address = replace(address,'Ã','A') where address like '%Ã%';
update public.register set address = replace(address,'Ä','A') where address like '%Ä%';
update public.register set address = replace(address,'Å','A') where address like '%Å%';
I want to declare arrays like this:
SET array1 = ARRAY['Á','Â','Ã','Ä','Å]
SET array2 = ARRAY['A','A','A','A','A']
and make a loop just like:
for(i=0; i<count(array1[]); i++){
update public.register set address = replace(address,array1[i],array2[i]) where address like '%array1[i]%';
}
I have not found any good and working idea on forums.
Thanks for any info about this :)

Answer 1

You don't need a loop. This can be done using a single UPDATE statement using regexp_replace()
update register
set address = regexp_replace(address, '[ÁÂÃÄÅ]', 'A', 'g')
where address ~ '[ÁÂÃÄÅ]';
'[ÁÂÃÄÅ]' matches each character in the list. The 'g' option will replace all occurrences in the value.
The WHERE clause then makes sure that only those rows that contain these characters in the address column are changed.
It would work without the WHERE as well, as the regexp_replace() will not change the address if the characters aren't contained. But it's good coding practice to only update those rows that do actually change.
Another option is to install the unaccent extension, then you can use:
update register
set address = unaccent(address)
where address <> unaccent(address);
This has the additional benefit, that you can use language specific rules if you want.