Oracle Version: Docker Image container-registry.oracle.com/database/enterprise:19.19.0.0
Setup Script (to enable logminer)

SELECT log_mode FROM v$database;
SHUTDOWN IMMEDIATE;
STARTUP MOUNT;
ALTER DATABASE ARCHIVELOG;
ALTER DATABASE OPEN;
SELECT log_mode FROM v$database;


ALTER DATABASE ADD SUPPLEMENTAL LOG DATA (PRIMARY KEY) COLUMNS;
select supplemental_log_data_min, supplemental_log_data_pk from v$database;

I'm using logMiner to get the redo logs, while i notice this strange behavior:

- Client Side Queries:

SELECT CURRENT_SCN FROM V$DATABASE; -- 2554427
INSERT INTO t1 VALUES (1, 1);
SAVEPOINT s1;
INSERT INTO t1 VALUES (2, 2);
SELECT CURRENT_SCN FROM V$DATABASE; -- 2554481
INSERT INTO t1 VALUES (3, 3);
SELECT CURRENT_SCN FROM V$DATABASE; -- 2554528
ROLLBACK TO s1;
SELECT CURRENT_SCN FROM V$DATABASE; -- 2554565
COMMIT;
SELECT CURRENT_SCN FROM V$DATABASE; -- 2554583

- To query logMiner for all the redo logs in this transaction:

EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo01.log', OPTIONS => DBMS_LOGMNR.NEW);
EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo02.log', OPTIONS => DBMS_LOGMNR.ADDFILE);
EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo03.log', OPTIONS => DBMS_LOGMNR.ADDFILE);
EXEC DBMS_LOGMNR.START_LOGMNR(STARTSCN => 2554427, ENDSCN => 2554565, OPTIONS => DBMS_LOGMNR.DICT_FROM_ONLINE_CATALOG);
SELECT SCN, XID, XIDUSN, XIDSLT, XIDSQN, SQL_REDO, ROLLBACK FROM V$LOGMNR_CONTENTS WHERE TABLE_NAME = 'T1';

This gives me:

       
SCN XID                    XIDUSN    XIDSLT    XIDSQN SQL_REDO                                                             ROLLBACK 
__________ ___________________ _________ _________ _________ _________________________________________________________________ ___________ 
   2554454 D4001F0014000000          212        31        20 insert into "MY_USER"."T1"("PK","V") values ('1','1');                      0 
   2554475 D4001F0014000000          212        31        20 insert into "MY_USER"."T1"("PK","V") values ('2','2');                      0 
   2554524 D4001F0014000000          212        31        20 insert into "MY_USER"."T1"("PK","V") values ('3','3');                      0 
   2554545 D4001F0014000000          212        31        20 delete from "MY_USER"."T1" where ROWID = 'AAASkaAAHAAAAFcAAC';              1 
   2554545 D4001F0014000000          212        31        20 delete from "MY_USER"."T1" where ROWID = 'AAASkaAAHAAAAFcAAB';              1 

- To query logMiner for only the ROLLBACK TO s1 stmt:

EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo01.log', OPTIONS => DBMS_LOGMNR.NEW);
EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo02.log', OPTIONS => DBMS_LOGMNR.ADDFILE);
EXEC DBMS_LOGMNR.ADD_LOGFILE(LOGFILENAME => '/opt/oracle/oradata/ORCLCDB/redo03.log', OPTIONS => DBMS_LOGMNR.ADDFILE);
EXEC DBMS_LOGMNR.START_LOGMNR(STARTSCN => 2554528, ENDSCN => 2554565, OPTIONS => DBMS_LOGMNR.DICT_FROM_ONLINE_CATALOG);
SELECT SCN, XID, XIDUSN, XIDSLT, XIDSQN, SQL_REDO, ROLLBACK FROM V$LOGMNR_CONTENTS WHERE TABLE_NAME = 'T1';

This gives:

       SCN XID                    XIDUSN    XIDSLT        XIDSQN SQL_REDO                                                             ROLLBACK 
__________ ___________________ _________ _________ _____________ _________________________________________________________________ ___________ 
   2554545 D4001F00FFFFFFFF          212        31    4294967295 delete from "MY_USER"."T1" where ROWID = 'AAASkaAAHAAAAFcAAC';              1 
   2554545 D4001F00FFFFFFFF          212        31    4294967295 delete from "MY_USER"."T1" where ROWID = 'AAASkaAAHAAAAFcAAB';              1 

So we can see that the actual XID should be `D4001F0014000000`. But if I narrow the SCN window for logMiner for just the ROLLBACK TO SAVEPOINT stmts, the last 8 digits of the XID will be padded with `FFFFFFFF`.

I'm wondering
- why this would happen?
- Is it possible to group sql_redo with the composite key of `(XIDUSN, XIDSLT)`, which seems to be consistent regardless of the scn window selection? In other words, if I found redo logs that share the same (XIDUSN, XIDSLT), can i say that they belong to the same transaction?