PostgreSQL, how to update an array into an existing array without creating duplicates?-CodePudding

I've a row with a column of text[] which looks like:

{"FName1 LName1", "FName3 LName3", "FName4 LName4"}

Next would I like to merge another array of names into that already stored array, but without creating any duplicates in it. How can I do that?

For example the other array I want to include looks like:

{"FName4 LName4", "FName1 LName1", "FName2 LName2"}

As a result would I like it to be stored like:

{"FName1 LName1", "FName2 LName2", "FName3 LName3", "FName4 LName4"}

The elements within the new array can be in any order, as long as no element (here names) are duplicated within that new (created) stored array.

I couldn't find an PostgreSQL array function which does that for me. I only figured out how to append or prepend an array (with duplicates).

Is there a way without having to read that row first? I really don't want to read the row first.

Edit based on the comment:

Create table including test row example:

CREATE TABLE users (
 indexid TEXT PRIMARY KEY,
 name_list TEXT[] NOT NULL
);
INSERT INTO users VALUES ('iid_123', '{"FName1 LName1", "FName3 LName3", "FName4 LName4"}')

Many thanks to the solution from @a_horse_with_no_name which now works for me like:

UPDATE users SET name_list=append_unique(name_list, '{"FName4 LName4", "FName1 LName1", "FName2 LName2"}') WHERE indexid = 'iid_123'

Documentations:

https://www.postgresql.org/docs/current/functions-array.html#ARRAY-FUNCTIONS-TABLE https://www.postgresql.org/docs/current/functions-string.html

CodePudding user response：

You will have to write a function to deal with that:

create function append_unique(p_input text[], p_to_append text[])
  returns text[]
as
$$
   select array_agg(element)
   from (
      select *
      from unnest(p_input) as t(element)
      union 
      select *
      from unnest(p_to_append) as t(element)
   ) x
$$
language sql
immutable;

The UNION will remove the duplicates.