recursive calls
Sujit Pandey, February 02, 2017 - 4:31 am UTC
recursive calls were 393. re-executing the same query returns 0 for recursive calls. I think i got my answer through some testing. Its through turning off the dynamic sampling which returns "consistent reads" close to the number of blocks for the table. Still my question is the "consistent reads" higher due to the recursive calls which hard parses the query? Is it due to the reading of oracle statistics tables for the parsing?
February 03, 2017 - 7:18 pm UTC
A dynamic sample is reading the tables *within the query* not (just) the dictionary statistics. So this could definitely account for increased logical io
recursive calls
Sujit Pandey, February 02, 2017 - 5:17 pm UTC
Recursive calls were 393. re-executing the same query returns 0 for recursive calls. I think i got my answer through some testing. Its through turning off the dynamic sampling which returns "consistent reads" close to the number of blocks for the table. Still my question is the "consistent reads" higher due to the recursive calls which hard parses the query? Is it due to the reading of oracle statistics tables for the parsing?
February 03, 2017 - 4:19 pm UTC
Yes, the hard parsing and reading the stats is part of this. If you're interested, trace the session. You can see what the recursive SQL statements are and how much work they do from this.
But..
A reader, February 04, 2017 - 6:54 am UTC
But in the example showed when creating table T, you didn't run any Stat. So how this be coherent with the last post?
Thanks Chris and Connor for your inputs
Sujit, February 06, 2017 - 7:02 am UTC
Thanks Chris and Connor for your inputs
February 06, 2017 - 2:45 pm UTC
thanks for the feedback