CREATE OR REPLACE FUNCTION schema_xyz.get_name_pg(my_uid_ bigint)
RETURNS character varying
LANGUAGE plpgsql
AS $function$
DECLARE
ergebnis varchar;
min_hi_date timestamp;
max_hi_date timestamp;
anz_namen bigint;
anz_namen_II bigint;
BEGIN
/*
We are looking for a name with a maximum first_date. If there is more than one, just take one...
*/
select count(distinct name_d) into anz_namen from schema_xyz.u_name where uid_=my_uid_;
IF (anz_namen=1)
THEN
select distinct(name_d) into ergebnis from schema_xyz.u_name where uid_=my_uid_;
ELSE
ergebnis:=null;
select min(first_date) into min_hi_date from schema_xyz.relation where uid_1=my_uid_ and uid_0=0;
select max(last_date) into max_hi_date from schema_xyz.relation where uid_1=my_uid_ and uid_0=0;
select count(name_d) into anz_namen_II
from schema_xyz.u_name
where uid_=my_uid_
and first_date=(SELECT max(first_date) from schema_xyz.u_name where uid_=my_uid_ and first_date>=min_hi_date and last_date<=max_hi_date);
if (anz_namen_II>0)
then
select name_d into ergebnis
from
(SELECT name_d
from schema_xyz.u_name
where uid_=my_uid_
and first_date=(select max(first_date) from schema_xyz.u_name where uid_=my_uid_ and first_date>=min_hi_date and last_date<=max_hi_date)) alias3 LIMIT 1;
end if;
-- if nothing is found in this way, you can still search for a name that completely covers the HI time
if (ergebnis is null)
then
select name_d into ergebnis
from (SELECT name_d
from schema_xyz.u_name
where uid_=my_uid_
and first_date<=min_hi_date and last_date >= max_hi_date) alias1 LIMIT 1;
end if;
-- Then there are still those with a name change in the middle. Simply add names with a maximum last_date and you're done.
if (ergebnis is null)
then
select name_d into ergebnis
from
(SELECT name_d
from schema_xyz.u_name
where uid_=my_uid_
and last_date=(select max(last_date) from schema_xyz.u_name where uid_=my_uid_)) alias3 LIMIT 1;
