1) as per the article referenced by Laurentiu Oprea
2) That's not the outline information, it's the Query Block / Object Alias
information (format =>'alias'). If you can find the outline information you
should be able to find the INDEX() - or index_rs_asc() hint for that access
and modify it to a suitable FULL() hint. (I assume this isn't the 2nd table
of a nested loop - or you might also need to find the USE_NL() and change
it to a use_hash() hint etc...)
If you can't get at the outline, there's a good change that the hint you
need is: FULL(@SEL$63682743 SRCL@SEL$10)
Bear in mind that if this makes the optimizer think the strategy it too
expensive it may "obey" your hint by doing the calculation and then choose
a complete different execution path that doesn't produce that query block,
thus making the hint unusable.
On Mon, 13 Sept 2021 at 18:17, Chris Taylor <
Background: I'm looking over Maria Colgan's posts here:
I have 2 questions:
1.) Can I assign "sql_text=>" to v_sql where I load v_sql from
dba_hist_sqltext for a sql_id?
I assume the answer should be yes?
2.) The patch I need to apply is in a terrible SUBSELECT thats part of a
In the execution plan it looks like this:
|* 43 | TABLE ACCESS BY GLOBAL INDEX ROWID BATCHED |
SPECIAL_REVERSED_CLAIM_LOG | 1 | 29 | 3 (0)| 00:00:01 |
ROWID | ROWID |
|* 44 | INDEX RANGE SCAN |
SPECIAL_REVERSED_CLAIM_LOG_IDX | 1 | | 2 (0)| 00:00:01 |
Which is this in the outline information:
43 - SEL$63682743 / SRCL@SEL$10
44 - SEL$63682743 / SRCL@SEL$10
So how I would hint this sql using the SQL Patch?
Would it be something like: