After creating index query is taking more

Question

Hello all,
I am trying the following..and not getting why the query is taking more time after creating the index
SQL> SELECT distinct version_no
2 FROM iefloa_data_anal
3 WHERE datetime_inserted > to_date('17-02-2004','dd-mm-yyyy');
VERSION_NO
999
Elapsed: 00:00:22.16
Execution Plan
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=15 Card=1 Bytes=22)
1 0 SORT (UNIQUE) (Cost=15 Card=1 Bytes=22)
2 1 TABLE ACCESS (FULL) OF 'IEFLOA_DATA_ANAL' (Cost=2 Card=1
Bytes=22)
/*******************Here I am creating the Index************/
SQL> create index t_ind_date on iefloa_data_anal(datetime_inserted);
Index created.
Elapsed: 00:07:427.38
SQL> SELECT distinct version_no
2 FROM iefloa_data_anal
3 WHERE datetime_inserted > to_date('17-02-2004','dd-mm-yyyy');
VERSION_NO
999
Elapsed: 00:00:40.10
Execution Plan
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=14 Card=1 Bytes=22)
1 0 SORT (UNIQUE) (Cost=14 Card=1 Bytes=22)
2 1 TABLE ACCESS (BY INDEX ROWID) OF 'IEFLOA_DATA_ANAL' (Cos
t=1 Card=1 Bytes=22)
3 2 INDEX (RANGE SCAN) OF 'T_IND_DATE' (NON-UNIQUE) (Cost=
1 Card=1)
can anybody explains whats the reason...
the table is having around 1.5 crore records
regards

Answer 1

The execution plan clearly shows that Oracle thinks only one row is going to match that condition on the index ("Card=1"). Is that a true statement? If not, you should make sure your stats are accurate. If the data is highly skewed, you may need to put a histogram on that column.
Another possibility is that the index is on much slower disks than the table. Although, I can't imagine any range scan on an index that has only one matching value taking that long. How about doing a full trace (SET AUTOTRACE ON) and posting the logical and physical i/o information?