SQL & PL/SQL :: Procedure To Update Double Subcategory Sequence

I need to write a procedure to be able to reuse it to clean up subcategory sequence. Here is a problem: two tables: Category and subcategory:

create table Category
(category_id number,
Name varchar2(20))
/
create table Sub_Category
(sub_category_id number,
category_id number,
sequence number)
/

In the ideal world each subcategory of a single category would have unique sequence so if there are 3 subcategory for the same category then each of them would have 1,2,3 in sequence, if there are 5 subcategories then 1,2,3,4,5 for each of them etc.

Problem I'm facing is that some of the subcategories sequences for the same category has the same values . For instance for 4 subcategories of the same category, each of them has 1 (1,1,1,1) in a sequence.

So ideal world is :

Insert into Category values (123 ,'Category1');
Insert into Category values (234 ,'Category2');
Insert into Category values (345 ,'Category3');
Insert into Category values (456 ,'Category4');
Insert into Category values (567 ,'Category5');
Insert into Sub_Category values (1,123,1);
Insert into Sub_Category values (2,123,2);
Insert into Sub_Category values (3,123,3);

But I've also bad rows like this:

Insert into Sub_Category values (4,234,1);
Insert into Sub_Category values (5,234,1);
Insert into Sub_Category values (6,234,1);
Insert into Sub_Category values (7,345,1);
Insert into Sub_Category values (8,345,1);
[code].....

Fix for this and my goal is to select all such cases where subcases have mixed up sequences as above and give them randomly numbers starting from 1. So if there are 3 subcategories like for CATEGORY 2 then just apply random number to the sequence of the subCATEGORIES like 1,2,3. For CATEGORY 3 : 1,2,3 to 7.

I was thinking to write two procedures one selecting all the categories and passing category ID to the other procedure that would actually update sequence, like this:

CREATE OR REPLACE PROCEDURE SCHEMA.SELECT_CATEGORY
IS
CURSOR c1
IS
select category_ID from category where ...;
BEGIN
FOR a IN c1 LOOP
UPDATE_SUBCATEGORY(a.Category_id);
COMMIT;
END LOOP;
END;
/

And the actual procedure updating subcategory:

CREATE OR REPLACE PROCEDURE SCHEMA.UPDATE_SUBCATEGORY
BEGIN
............
END;
/

write PROCEDURE SCHEMA.SELECT_CATEGORY cursor to not miss any of the categories ID having mixed up subcategory. There can be any of the doubled sequences like doubled 1 value (this is majority) but there can be any other doubles (or at least I need to make sure that there aren't any other doubles 2 values or 3 values in sequence etc.)

And how to write SCHEMA.UPDATE_SUBCATEGORY to loop through rows of subcategory and update sequence with values starting from 1 ?