Help explain plan

From: Peter Laursen <ptl_at_edbgruppen.dk>
Date: 8 Jun 1999 10:47:16 GMT
Message-ID: <01beb1a5$113070a0$2c289a0a@apollo>

Could someone help me understand this execution plan? Is the index FK1_SAMPLE actually used?
What could I do to optimize this query?

SELECT SUM(DECODE(M.METERACTION,1,S.VALUE,-S.VALUE)) VALUE, S.SAMPLETIME FROM METER M, LASTSAMPLE S
WHERE M.METEREDI_RECNUM=:HOST1 AND M.METERACTION<>0 AND M.METER_RECNUM=S.METER_RECNUM
AND S.SAMPLETIME BETWEEN :HOST2 AND :HOST3 GROUP BY S.SAMPLETIME CREATE INDEX FK1_SAMPLE ON
  SAMPLE(METER_RECNUM, SAMPLETIME) ... SELECT STATEMENT Optimizer=RULE
  SORT (GROUP BY)
    MERGE JOIN (OUTER)

      SORT (JOIN)
        MERGE JOIN (OUTER)
          SORT (JOIN)
            NESTED LOOPS
              TABLE ACCESS (FULL) OF SAMPLE
              TABLE ACCESS (BY INDEX ROWID) OF METER
                INDEX (UNIQUE SCAN) OF PK_METER (UNIQUE)
          SORT (JOIN)
            TABLE ACCESS (FULL) OF SAMPLE_QUALITY
      SORT (JOIN)
        TABLE ACCESS (FULL) OF SAMPLE_USER

Thanks
Peter Received on Tue Jun 08 1999 - 05:47:16 CDT