Re: bloom filter not used after 19C upgrade
Date: Fri, 10 May 2024 09:31:07 +0200
Message-ID: <b21ec264-3269-4249-9754-e0710be6272c_at_bluewin.ch>
So your point is that since bloom filters are determined on runtime, but
10053 is created on parse time. Thus a 10053 can not contain complete
information on bloom filter usage.
A point well taken. The reference to explain plan is misleading. The
point is about parsing, not explain plan .
However, the decision for a plan can determine if bloom filters can be
used in principle.
There are plans to prevent that.
Thanks
Lothar
Am 09.05.2024 um 14:36 schrieb Mladen Gogala:
> On Thu, 2024-05-09 at 08:30 +0200, Lothar Flatz wrote:
>> Malden,
>>
>> explain plan is out of question. Please stop wasting my time by
>> mentioning it.
>
> I mentioned explain plan because Laurentiu mentioned the event 10053.
> The event 10053 traces oracle optimizer. The SQL executor, which is
> separate from the CBO, can decide on its own to do or not to do Bloom
> filter. Consequently, the event 10053 can never show Bloom filters.
> --
> Mladen Gogala
> Database SME
> https://dbwhisperer.wordpress.com
>
-- http://www.freelists.org/webpage/oracle-lReceived on Fri May 10 2024 - 09:31:07 CEST