Topic: dynamic_sampling

Somebody somewhere came across notes with really good and deep explanation as works _?
At us on some request  spontaneously includes dynamic_sampling level = 5, thus for not clear reason is mistaken in  in tens and hundreds times though the analysis of 128 casual units in this case should be more than enough for a true estimation.
Whether it is possible that the matter is that a label  , and almost all necessary lines for the given request lie in the last partition (end_time = 2100 year), and he selects for the analysis casual partitions from 1048575 virtual partitions (though real there only some tens)?


Re: dynamic_sampling

https://blogs.oracle.com/optimizer/dyna … -optimizer
https://blogs.oracle.com/optimizer/opti … -release-2
http://oracle-randolf.blogspot.de/2009/ … ables.html

Valergrad wrote:

  spontaneously includes dynamic_sampling level = 5

oracle what version? And that there is an essential difference. Though I suspect that at you simply parallel performance and optimizer_dynamic_sampling :
[quote =] From Oracle Database 11g Release 2 onwards, the optimizer automatically decides if dynamic sampling will be useful and what dynamic sampling level will be used for SQL statements executed in parallel. This decision is based on size of the tables in the statement and the complexity of the predicates.

. If my memory doesn't fail me that is a difference between  2 or installed 2 (i.e. it is installed through alter system or a file of parameters): at installed manually 2 optimizer_dynamic_sampling at a parallel does not settle up automatically, and takes installed 2