Assigning an identifier

Question

I have the following table TEST3 with data as shown below.
I need to identify the records that have the same XLS and LORENTION values, but under one condition.
The condition is that if you look at the data ordered by seq asc, the records where seq in
(148206, 148207) should have a different identifier than the record where seq = 148215, although they all
have the same XLS and LORENTION values. The reason is because they are separated by a different XLS and LORENTION
set of records (seq in (148211, 148212, 148213, 1482114)).
In other words, I need a way to read record by record in ascending order of SEQ, assign an identifier for the same XLS and LORENTION values until it hits a different set of XLS and LORENTION values.
Any help will be much appreciated.
CREATE TABLE TEST3
SEQ NUMBER(28,7),
XLS VARCHAR2(255 BYTE),
CANS VARCHAR2(255 BYTE),
LORENTION VARCHAR2(255 BYTE),
COLOR VARCHAR2(255 BYTE)
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148206, 'Roucotine -Rs', 'Atscks, Sams', 'terss', 'required');
Insert into test3
(SEQ, XLS, CANS, lorention)
Values
(148207, 'Roucotine -Rs', 'Atscks, Polaris ', 'terss');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148207, 'Roucotine -Rs', 'Atscks, Polaris ', 'terss', 'yellow');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148207, 'Roucotine -Rs', 'Atscks, Polaris ', 'terss', 'dark red');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148207, 'Roucotine -Rs', 'Atscks, Polaris ', 'terss', 'red');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148211, 'Roucotine -Rs', 'Gdse', 'tebbbbbb', 'not used');
Insert into test3
(SEQ, XLS, CANS, lorention)
Values
(148212, 'Roucotine -Rs', 'Gdse, ee', 'tebbbbbb');
Insert into test3
(SEQ, XLS, CANS, lorention)
Values
(148213, 'Roucotine -Rs', 'Gdse, Right', 'tebbbbbb');
Insert into test3
(SEQ, XLS, CANS, lorention)
Values
(148214, 'Roucotine -Rs', 'Gland, rrt5', 'tebbbbbb');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148215, 'Roucotine -Rs', 'Rews', 'terss', 'required');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148215, 'Mac', 'Rews', 'terss', 'dark red');
Insert into test3
(SEQ, XLS, CANS, lorention, COLOR)
Values
(148215, 'Roucotine -Rs', 'Rews', 'terss', 'red');

Answer 1

version: Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bi
This is the source:
select seq, xls, lorention, cans, color from test3 order by seq;
I need something like:
select seq, xls, lorention, cans, color, max(seq) over (partition by xls, lorention) id from test3 order by seq;
The first 5 records should have one id because the xls and lorention are the same, then the next record, since lorention value changes, it should have a different id for rows 6 - 9. row 10 should have a different id because lorention value changes. Row 11 should have a different id because xls value changes. row 12 should have a different ID because xls changes. If you look at it, row 10 and row 12 have the same xls and lorention values as rows 1-5, but they need a different id because there are other xls and lorention values in between. In the example above, I used max(seq) as the identifier, but this won't work because rows 10 -12 should have different id but have the same seq values.