Question

Say I have a table, created as follows:

 CREATE TABLE test_table (id serial, unique_id varchar(50) primary key, name varchar(50));

 test_table
 ----------
 id | unique_id | name

In that table, I would like to update the unique_id field with the newly inserted id concatenated with the inserted name in a single go.

Usually this is accomplished by two queries. (PHP way)

 $q = "INSERT INTO table (unique_id,name) values ('uid','abc') returning id||name as unique_id;";  
 $r = pg_query($dbconn,$q);
 $row = pg_fetch_array($r);

 $q1 = "UPDATE test_table set unique_id =".$row['unique_id']." where unique_id='uid'"; 
 $r1 = pg_query($dbconn,$q1);

Is there any way to do the above in a single query?

Answer 1

You can have several options here, you could create a AFTER trigger which uses the generated ID for an direct update of the same row:

CREATE TRIGGER test_table_insert ON AFTER INSERT ON test_table FOR EACH ROW EXECUTE PROCEDURE test_table_insert();

And in your function you update the value:

CREATE FUNCTION test_table_insert() RETURNS TRIGGER AS $$
BEGIN
  UPDATE test_table SET uniqid = NEW.id::text || NEW.name WHERE id = NEW.id;
END;
$$ LANGUAGE plpgsql;

You need to add the function before the trigger.

An other option would be to do it directly in the insert:

INSERT INTO table (id, unique_id, name) values (nextval('test_table_id_seq'), 'abc', currval('test_table_id_seq')::text || 'abc') returning id;

But as a_horse_with_no_name pointed out, I think you may have a problem in your database design.