Creating multiple updates on database with procedure. Problem with incrementation - sql

Question

I'm trying to create a procedure in oracle sql in order to make multiple updates for one table. I have around 500 updates for different values but I'm only changing one column which is unique and it's task is to number rows. So i want to change number 20 to 21 and so on until number 120. So in the end number 20 will be empty and I will have 121 rows.
I created this loop but I keep getting error "unique constraint violated" so i figured that my iteration is not working.
declare
n integer := 20;
BEGIN
FOR x in 20..120 LOOP
Update my_table
SET number_of_field = n WHERE number_of_field > 20;
n:= n + 1;
END LOOP;
END;
I would really appreciate help because it's my first time using procedures in SQL and I will feel like I'm close to the solution.
Thank you!

Answer 1

Oracle is one of the few databases (PostgreSQL being the other one afaik) that allows you to defer constraint checks until the end of the transaction.
If you place the whole block of code into a single transaction, and also defer the check of the UNIQUE constraint to DEFERRABLE INITIALLY DEFERRED, then you can perform all updates in it, even if you temporarily store duplicate values in the column. However, once you perform the COMMIT, the constraint (for all rows/updates) will be validated, and if it fails the transaction will be rolled back. So you need to be on a good stance once you issue the commit.
For example you can change your table to something like:
create table my_table (
... other_columns here
number_of_field,
constraint uq1 unique(number_of_field) deferrable initially deferred
);

Answer 2

This is how I understood the question:
SQL> -- Create table
SQL> create table my_table (number_of_field number, somecol varchar2(10));
Table created.
SQL> -- Insert sample data; my end with 30, not 120 (that's too much for an example)
SQL> insert into my_table
2 select 20 + level - 1, 'X'
3 from dual
4 connect by level <= 11;
11 rows created.
SQL> -- Table contents
SQL> select number_of_field, somecol from my_table order by number_of_field nulls first;
NUMBER_OF_FIELD SOMECOL
--------------- ----------
20 X
21 X
22 X
23 X
24 X
25 X
26 X
27 X
28 X
29 X
30 X
11 rows selected.
Two-step job: update existing rows (by adding 1) and insert an empty one (somecol is here just to show that there's a row with empty number_of_Field column).
SQL> -- Update existing rows; insert an "empty" one
SQL> begin
2 update my_table set
3 number_of_field = number_of_field + 1;
4 insert into my_table (number_of_field, somecol)
5 values (null, 'X');
6 end;
7 /
PL/SQL procedure successfully completed.
SQL> -- The result
SQL> select number_of_field, somecol from my_table order by number_of_field nulls first;
NUMBER_OF_FIELD SOMECOL
--------------- ----------
X
21 X
22 X
23 X
24 X
25 X
26 X
27 X
28 X
29 X
30 X
31 X
12 rows selected.
SQL>