On 11/25/2011 10:52 AM, Jeff Janes wrote:
Given Peter's patch on the same subject, should we now mark this one
as rejected in the commitfest app?
https://commitfest.postgresql.org/action/patch_view?id=681
That may be premature. While the testing I've done so far suggests
Peter's idea is the
On Sat, Nov 5, 2011 at 8:42 PM, Tomas Vondra t...@fuzzy.cz wrote:
Dne 6.11.2011 03:16, Peter Geoghegan napsal(a):
2011/11/6 Tomas Vondra t...@fuzzy.cz:
Hi everyone,
The patch implements a simple cleaning that replaces the parameter
values with generic strings - e.g. numbers are turned to :n,
On Sun, Nov 6, 2011 at 2:56 PM, Greg Smith g...@2ndquadrant.com wrote:
Peter's patch adds a planner hook and hashes at that level. Since this cat
is rapidly escaping its bag and impacting other contributors, we might as
well share the work in progress early if anyone has a burning desire to
On 11/07/2011 09:03 AM, Robert Haas wrote:
I think it's an established principle that the design for features
like this should, for best results, be discussed on -hackers before
writing a lot of code.
You can see from the commit history this idea is less than a month old.
Do we need to get
On 11/06/2011 06:00 PM, Tom Lane wrote:
Peter Geogheganpe...@2ndquadrant.com writes:
A major consideration was backwards compatibility;
This is not a consideration that the community is likely to weigh
heavily, or indeed at all. We aren't going to back-port this feature
into prior release
On Mon, Nov 7, 2011 at 10:27 AM, Greg Smith g...@2ndquadrant.com wrote:
On 11/07/2011 09:03 AM, Robert Haas wrote:
I think it's an established principle that the design for features
like this should, for best results, be discussed on -hackers before
writing a lot of code.
You can see from
Peter Geoghegan pe...@2ndquadrant.com writes:
I'm a couple of days away from posting a much better principled
implementation of pg_stat_statements normalisation. To normalise, we
perform a selective serialisation of the query tree, which is hashed.
That seems like an interesting approach, and
Tomas Vondra t...@fuzzy.cz writes:
If it was possible to compare the actual plan (or at least a hash of the
plan), and keeping one record for each plan, that'd be extremely nice.
You'd see that the query was executed with 3 different plans, number of
calls, average duration etc.
I like that
On 6 Listopad 2011, 16:08, Tom Lane wrote:
Peter Geoghegan pe...@2ndquadrant.com writes:
I'm a couple of days away from posting a much better principled
implementation of pg_stat_statements normalisation. To normalise, we
perform a selective serialisation of the query tree, which is hashed.
On 11/06/2011 01:08 PM, Tom Lane wrote:
Peter Geogheganpe...@2ndquadrant.com writes:
... It also does things
like intelligently distinguishing between queries with different
limit/offset constant values, as these constants are deemed to be
differentiators of queries for our purposes. A
On 6 November 2011 15:08, Tom Lane t...@sss.pgh.pa.us wrote:
Are you intending to hash the raw
grammar output tree, the parse analysis result, the rewriter result,
or the actual plan tree?
I'm hashing the Query tree from a planner plugin (function assigned to
planner_hook), immediately prior
Peter Geoghegan pe...@2ndquadrant.com writes:
On 6 November 2011 15:08, Tom Lane t...@sss.pgh.pa.us wrote:
Are you intending to hash the raw
grammar output tree, the parse analysis result, the rewriter result,
or the actual plan tree?
I'm hashing the Query tree from a planner plugin
2011/11/6 Tomas Vondra t...@fuzzy.cz:
Hi everyone,
The patch implements a simple cleaning that replaces the parameter
values with generic strings - e.g. numbers are turned to :n, so the
queries mentioned above are turned to
SELECT abalance FROM pgbench_accounts WHERE aid = :n
and thus
Dne 6.11.2011 03:16, Peter Geoghegan napsal(a):
2011/11/6 Tomas Vondra t...@fuzzy.cz:
Hi everyone,
The patch implements a simple cleaning that replaces the parameter
values with generic strings - e.g. numbers are turned to :n, so the
queries mentioned above are turned to
SELECT abalance
14 matches
Mail list logo