Oracle: delete duplicate rows from a table, to have only one row for each unique expression

?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

DELETE FROM EXTERNAL_ENTITIES WHERE    ROWID IN    (       SELECT "ROWID" -- ESCAPED, BECAUSE WE NEED THE INNER ROWID, NOT THIS TMPTABLE'S ROWID       FROM          (             SELECT                RANK() OVER(PARTITION BY NAME ORDER BY ROWID) RANK_N, -- ORDER IN THE SUBGROUP                ROWID AS "ROWID"                                    , -- ESCAPE THE ROWID             FROM EXTERNAL_ENTITIES             WHERE                NAME IN                (                   SELECT NAME                   FROM EXTERNAL_ENTITIES                   GROUP BY NAME                   HAVING COUNT(*) > 1 -- ONLY DUPLICATES OR MORE                )          )       WHERE RANK_N > 1    );

or

?

1 2 3 4 5 6 7 8

DELETE FROM CUSTOMFIELDVALUE WHERE ROWID IN    (SELECT T."ROWID" FROM       (SELECT RANK() OVER(PARTITION BY ISSUE, CUSTOMFIELD ORDER BY ROWID) RANK_N, ROWID AS "ROWID"          FROM CUSTOMFIELDVALUE          WHERE CUSTOMFIELD IN (10180, 10216, 10336, 10340, 10464, 10550, 10590, 10911, 14714) -- customfield ID-s come here, taken from the log.       ) T    WHERE T.RANK_N > 1);