How to replace multiple special characters in Postgres 9.5

replace()

If you want just to replace one or few characters you can use function replace(string text, from text, to text) that replaces all occurrences in string substring. The replace function can be used to replace one character to several characters.

translate()

If you want to translate some letters to other letters you can user function translate(string text, from text, to text) that replaces any character in a string that matches a character in the from by the corresponding character in the to set.

Some data to play with:

drop table if exists xyz;

create table xyz (
    id serial not null,
    name varchar(30)
);

insert into xyz (name) values
    ('Juhänäo'),
    ('Jürgüen'),
    ('Dannäu'),
    ('Übüdyr');

Example of replace function:

select replace(name, 'ä', 'a') from xyz;

This function replaces letter ä in the name column with letter a. Juhänäo becomes Juhanao.

select replace(name, 'ä', 'ae') from xyz;

Now it replaces letter ä with ae.

select replace(replace(replace(name, 'ä', 'ae'), 'ü', 'ue'), 'Ü', 'Ue') from xyz;

Not very nice, but in the example all ä become ae, ü become ue, and Ü become 'Ue'.

update xyz set name = replace(replace(replace(name, 'ä', 'ae'), 'ü', 'ue'), 'Ü', 'Ue');

Changes letters and updates rows. The result of the update is following:

Juhaenaeo
Juergueen
Dannaeu
Uebuedyr

Example of translate function:

select translate(name, 'ä,ü,Ü', 'a,u,U') from xyz;

Translates all letters ä to a, ü to u and Ü to U.

update xyz set name = translate(name, 'ä,ü,Ü', 'a,u,U');

Updates table so all predefined letters are translated and the change is saved to the database. The result of the update is following:

Juhanao
Jurguen
Dannau
Ubudyr

More information:

Replace characters with multi-character strings

Postgresql string functions

Avoid writing your own unaccent function. Instead, I recommend using unaccent extension.

create extension unaccent;
select unaccent('Juhänäo');
select unaccent(name) from xyz;

Here is a link to Postgres documentation for Unaccent

you simply use this :

select  translate(column, 'âàãáéêèíóôõüúç', 'aaaaeeeiooouuc') from table

Good luck ^^

If you are after German letters, then this works:

CREATE OR REPLACE FUNCTION public.udf_transliterate_german(
    german_word character varying)
    RETURNS character varying
    LANGUAGE 'sql'
    COST 100
    IMMUTABLE PARALLEL UNSAFE
AS $BODY$
SELECT REPLACE(
    REPLACE(
        REPLACE(
            REPLACE(
                REPLACE(
                    REPLACE(
                        REPLACE(
                            REPLACE(german_word,
                                'ä','ae'),
                            'ö','oe' ),
                        'ü','ue'),
                    'ß','ss'),
                'Ä', 'AE'),
            'Ö', 'OE'),
        'Ü', 'UE'),
    'ẞ', 'SS');
$BODY$;

It is not elegant though.

This is a generic recursive function written in python (but can easily be replicated in whatever language you prefer) That takes in the original string and a list of substrings to be removed:

def replace_psql_string(str_to_remove, query):

    ss = str_to_remove.split(',')
    if ss[0] != '':
        query = "REGEXP_REPLACE('{0}', '{1}', '', 'gi')".format(query, ss[0])
        return self.replace_psql_string(','.join(ss[1:]), query)
    else:
        return query

# Run it
replace_psql_string("test,foo", "input test string")

The code splits the comma separated string to a list of substrings to remove, Creates a REGEXP_REPLACE function in sql with the first substring. Then, recursively calls the function again with [1:] elements and append each REGEXP query to the passed parameter. So that finally, when no more substrings are in the list to remove, the final aggregated query is returned.

This will output:

REGEXP_REPLACE(REGEXP_REPLACE('input test string, 'test', '', 'gi'), 'foo', '', 'gi')

Then use it in whatever part of your query that makes sense. Can be tested by appending a SELECT in front:

SELECT REGEXP_REPLACE(REGEXP_REPLACE('input test string, 'test', '', 'gi'), 'foo', '', 'gi');


regexp_replace 
----------------
 input  string
(1 row)