Here is a patch submission for reference to the next commitfest.
Improve pgbench measurements progress report
- Use progress option both under init bench.
Activate progress report by default, every 5 seconds.
When initializing, --quiet reverts to the old every 100,000 insertions
Split 2 of the initial submission
pgbench: reduce and compensate throttling underestimation bias.
This is a consequence of relying on an integer random generator,
which allow to ensure that delays inserted stay reasonably in
range of the target average delay.
The bias was about 0.5% with 1000
Split 3 of the initial submission, which actually deal with data measured
and reported on stderr under various options.
This version currently takes into account many comments by Noah Mish. In
particular, the default no report behavior under benchmarking is not
changed, although I really
, but with much more impact on the code, would be to
be able to change the size at database/table level.
Any thoughts?
--
Fabien.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
This patch is pretty trivial.
Add modulo operator to pgbench.
This is useful to compute a permutation for tests with non uniform
accesses (exponential or gaussian), so as to avoid trivial correlations
between neighbour keys.
--
Fabien.diff --git a/contrib/pgbench/pgbench.c
me kind of cypher
function but which does not operate on power-of-two sizes...
* if all necessary features are available, being able to run a strict
tpc-b bench (this mean adapting the init phase, and creating a
new builtin script which matches the real spec): no big deal.
Any thoughts on th
This minor patch shows the expected drawing percent in multi-script
reports, next to the relative weight.
--
Fabiendiff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index 4196b0e..3b63d69 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -3080,10
Remove pgbench clientDone unused "ok" parameter.
I cannot get the point of keeping a useless parameter, which is probably
there because at some point in the past it was used. If it is needed some
day it can always be reinserted.
--
Fabien.diff --git a/src/bin/pgbench/pgbench.c
, as the
INSTEAD keywords which tells that it is not. I also think the syntax
would be more balanced with the keyword.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http
, as the
INSTEAD keywords which tells that it is not.
--
Fabien COELHO _ [EMAIL PROTECTED] _ http://www.coelho.net
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs
in pg_restore.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
(quand TIMESTAMP REFERENCES day, ...);
The intent could be that events should refer to some day already
registered in the base. Obviously it does work, because the = will cast to
timestamp, to only the 00:00:00 timestamp will match a day.
etc.
--
Fabien Coelho - [EMAIL PROTECTED
and 2.
Sure, I agree with you there are example where you may mean that.
My point is about a ***warning***, because I think that most of the time
this hides a future bug, even if some time it may be the intent.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED
declared a foreign key on your table,
so you already get the warning at that time. If you chose to
ignore it, you must be right!
3/ If you bother to declare foreign keys, you want them to be sound,
so if postgres can help a little bit, that should not harm.
--
Fabien Coelho - [EMAIL
to the TODO list,
and I wish to claim it or part of it.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL
- processing constraint $1: CHECK (ind0)
-
- processing constraint foo: xxx REFERENCES bye
-
- processing constraint bla: data NOT NULL
-
-
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 1: subscribe
the result looks good,
and are most of the time useful.
stmt: CREATE TABLE ( anything ); { generate error missing table name }
Yes, that's an idea, but it would change the syntax definition, and
may create conflicts, so it's an harder way to do it.
--
Fabien Coelho - [EMAIL PROTECTED
Dear patchers,
New submission attempt, with new message wording. Also, the message is
only shown once, and the offending key component is not shown.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
costly_ri_notice.patch.gz
Description: Binary data
---(end
more appealing to me;-)
Or am still I off my shoes? (translator: a cote de mes pompes;-);
Thanks, for any advice on this point or others,
have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget
an integer array of character start
offsets and then work with that as if it were a wchar array.
Yep, I thought about this kind of thing.
Thanks for your advice,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9
Dear patchers,
Sorry, wrong list:-(
--
Fabien.
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send unregister YourEmailAddressHere to [EMAIL PROTECTED])
for your comments and ideas, I'll submit a v3 on (my) tomorrow.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
it.
Well, if it is the only issue against the adoption of the patch, I will do
without the line number.
I'll submit a new patch in a moment.
Thanks for your comments,
--
Fabien.
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please
an integer value;-) I have to convert to a string, or
deal directly with the line number.
With my new current implementation, the line number is only shown if the
query is more than 3 lines.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
are at risk with the current scheme?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
Fabien COELHO wrote:
There is also a localisation issue here, as the translation of both
lines must match so that the alignment is kept. I thought that if it
is the very same word, the translation should be the same.
You can just indent with as many spaces. This is done in other places
PQmblen.
Ok, what you describe is basically what I've done with the qidx
computation as suggested by Tom Lane and then later I check that the
encoded length is one to find my special characters.
Thanks for you reply,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
that could
be improved later on...
--
Fabien COELHO.
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
On Sat, 13 Mar 2004, Tatsuo Ishii wrote:
Oops. I have included this time.
How ! a japanese vi !
I could write PQmbtermlen function for every encoding supported by
PostgreSQL
That would be great ! ;-)
Ok, I will work on this.
Thanks.
--
Fabien Coelho - [EMAIL PROTECTED
, but I would suggest the option to be on by default.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
looks misleading to me. I would take a
performance_advices or performance_checks as these are not used yet,
and the levels may be notice/warning... and the hint field is not
necessarily used.
Well, this is just to talk;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end
point a view, I think it would be great to have
some place to put warnings or advices, so even if it is outside, it is
better than nowhere;-)
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 3: if posting/reading
a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
in a relationnal database, so let us stay there;-)
Good night,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
pg_catalog, but I have not found yet what I was looking
for.
If you do not like some advices, just DROP VIEW the_advice;
If you do not like advices at all, just DROP SCHEMA xpg_advisor;
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]--
-- $Id: pg_advisor.sql,v 1.3 2004/03/19 10:41:47
.
The more I think about it, the more I find it should be the way to go,
rather than having a new external tool.
--
Fabien Coelho - [EMAIL PROTECTED]--
-- $Id: pg_advisor.sql,v 1.13 2004/03/19 14:55:39 coelho Exp $
--
-- pg_advisor maybe future system schema?
--
-- at the time it is xpg_advisor
for any pointer,
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
is alas necessary. I thought about a
P 22 SELECT something FRM table
but psql simply uses the raw P field content in the message, so you
would get:
... at or near character 22 SELECT something FRM table
with old clients:-(
--
Fabien Coelho - [EMAIL PROTECTED]
---(end
. The information is available, that is the point.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
) possible inclusion in postgresql?
- among other contributions? what about contrib/advisor?
- added to template1 on default installation?
maybe not for a first release? or yes? it is easier to communicate
about
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]--
-- $Id
?
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
(ARRAY['%a%','%o%']);
IMHO, it would make sense to allow the later unless some critical parser
implication, as an average user (say, me;-) would try that before the
first one.
= Add to todo list?
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
anyway.
Why not.
How much work in the infrastructure ?
What would be the added value ? better communication ?
Having the schema included in contrib/ might help adoption,
but so would pgadmin/phpgadmin.
Sure.
Thanks a lot for your comments. I'll submitted an updated version later.
--
Fabien
haven't looked at it yet.
As for what is not foreseen yet, who knows? ;-)
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send unregister
on undecidable elements
have a lower grade. For instance, most attributes should be NOT NULL
from a statistical point of view, but it is perfectly legitimate to
have nullable attributes mostly anywhere, so the corresponding advices
is just an info.
--
Fabien Coelho - [EMAIL PROTECTED
} Postgres
[JOKE] What about renaming postgreSQL myPostgres? ;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 8: explain analyze is your friend
for these explanations? Just like
in psql \* shortcuts query about pg_tables to show informations.
--
Fabien.
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
that outer-level context hooks should be able
to editorialize on what inner-level ones said (or perhaps vice versa?)
No.
I just missed the error_context_stack by focussing on errcontext(),
which does not managed a stack at all, so I was quite puzzled. Now
it is clearer.
Thanks,
--
Fabien Coelho
to compare the two, I have to resort to some hacks.
It also appears that there are very few available function to deal with
int2vector at the SQL or PL/pgSQL level. Basically I found none:-(
Any comments?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
or pg_constraint or
pg_index?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
,
but I cannot seen any major reason why this could not and should not be
done...
So my question is : would it be reasonnable to submit a new posgres.bki
version, by dropping some without_oids in some declarations??
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
for an
external tool?
It is sure possible to circumvent the issue by putting the needed
composite keys here and there, but simple plain oids would look better.
One concept/one field looks nicer.
Thanks in advance,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
required its rows to have unique OIDs, the
probability of collisions after OID-counter wraparound would be much
greater than it is in other catalogs.
Mmh. Maybe you could have considered sequences.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
. Are there others we
missed?
Sure:-)
Date: Wed, 17 Mar 2004 09:09:40 +0100 (CET)
From: Fabien COELHO [EMAIL PROTECTED]
To: PostgreSQL Patches [EMAIL PROTECTED]
Subject: [PATCHES] hint infrastructure setup (v3)
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED
Is it better in /contrib or gborg?
Gborg imho. I thought we were trying to move all non-core code there
now. Isn't that why psqlodbc etc. were moved?
The argument was that it can be devopped and released independently?
Features in contrib/ have a premium over external add-ons.
--
Fabien
say Silly, thats on gborg! and they look at me strangely
Sure. The gborg name does not strike as being related to postgresql.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose
?
Thanks in advance,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send unregister YourEmailAddressHere to [EMAIL PROTECTED])
to the
database.
This is not the actual usage I have in mind, but this could be a possible
usage for such a view.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http
.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 8: explain analyze is your friend
it must be dealt with in the backend. I would prefer
a backend solution, so that == is also always =.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
joining
. Either you're cool and homogeneous, or either
you're strict. At the time, postgreSQL is rather cool in some place:
!= and , ~~ and LIKE... and a little bit inhomogeneous.
--
Fabien.
---(end of broadcast)---
TIP 2: you can get off all lists at once
, say for AND, I'm not
sure the optimiser will know that is a AND and take that into account.
That's my view.
Yep.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
they will have to face if the
database they use is different. So why bother?
That's just my view.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send unregister
.
But now is too late;-)
My point is more practical: my students know one syntax, I don't have a
lot of hours for the course, most of them won't be software developers but
rather software users, so if I can do more interesting things it is not
bad.
--
Fabien Coelho - [EMAIL PROTECTED
be possible to mimic (theoretically) other SQL implementations.
With extended, I could have c-like operators;-) : ! || == !=
Still dreaming.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget to increase your free space
non portable code... But I'm not sure it will be
accepted;-)
Moreover, having == as a synonym for = is not necessarily in contradiction
with a stable, reliable and predictable server.
AND is not an operator, ...
The docs says logical operator. Maybe you mean is not a pg_operator.
--
Fabien
,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
skill.
for this purpose, I could have java-programming, and java details
are really important in this course. int and long differ, although in
C int and long may or may not differ.
That will serve them differently.
Have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED
mean no acl, thus only super users would
be able to access such objects?
Or am I missing something on this point?? That's quite possible.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget to increase your free space
,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
by default, for now */
^^^
world_default = ACL_EXECUTE;
...
case ACL_OBJECT_LANGUAGE:
/* Grant USAGE by default, for now */
world_default = ACL_USAGE;
The patch would be quick;-)
Have a nice day,
--
Fabien COELHO _ http
.
I'm not planing my students to know what i=++i+i++; could mean.
If I teach about java/c/c++/java, this may be an issue.
So it depends on the course goal.
Well, I'm happy that so many people have ideas about what to teach and how
to teach it;-)
Have a nice day,
--
Fabien Coelho - [EMAIL
-be-obvious hidden
rationnal behind this feature:
What editor or terminal is supposed to be able to generate text in
different encodings depending on the part of the sentence? I don't think I
have that in emacs. Or is it irrelevant??
I cannot see where I could use such a feature.
--
Fabien Coelho
to take care of conversions.
printf(SELECT * FROM foo WHERE field1 = '%s',
latin1_to_database_encoding(...));
Well, so this is a great new useful feature indeed, that will help improve
the lexer code a lot;-)
Good luck,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end
3) Implement and submit a patch
Any advice is greatly appreciated. Especially the ones that might speed up
step #1.
Comments, suggestions?
Good luck.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 5: Have you checked our
to Thomas that he should only touch
the search array to insert new pointers to guc description structures,
which may come from outside or be newly allocated, without touching the
existing static versions for internal guc variables.
--
Fabien Coelho - [EMAIL PROTECTED
psql SHOW TRANSACTION ISOLATION LEVEL;
transaction_isolation
---
read committed
I would have expected 'serializable' ?
If the isolation level or read/write mode is specified, the new
transaction has those characteristics, as if SET TRANSACTION was
executed.
--
Fabien Coelho
AS ON UPDATE TO foo
WHERE old.locked=TRUE
DO INSTEAD ERROR;
I think this simple new rule action could be added to pg.
I'm planning to do it, if there is no opposition.
Any comments on this proposed new rule action?
Or did I missed something obvious with rules?
Have a nice day,
--
Fabien
this sort of thing, and got burned by the fundamental
semantic differences ...
Ok, I'll look into that. Thanks for the pointer.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives
against a database that some
say is good enough (MySQL), and another one that some
say is too much (Oracle)
free and serious.
o Are our priorities too technically driven?
Not bad if other agendas can also get through.
--
Fabien Coelho - [EMAIL PROTECTED
/postgresql/server
ls: /usr/local/pg750a/include/postgresql/server: No such file or directory
Too bad, server headers are not installed by default:-(
--
Fabien.
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
statistics will have to be on by default
as well.
Good luck;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so
be an accessor;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 8: explain analyze is your friend
shortdescr... to specify the object name, its type and its description,
so that some filter could derive the appropriate COMMENT ON.
I appreciate having something to read on a \d+ in psql, without having
to turn to the online doc.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end
is reset as expected?
Thanks in advance for your answers and comments.
Have a nice day.
--
Fabien Coelho - [EMAIL PROTECTED]*** ./doc/src/sgml/catalogs.sgml.orig Mon Apr 5 12:06:40 2004
--- ./doc/src/sgml/catalogs.sgmlFri May 7 08:46:38 2004
***
*** 1536,1541
--- 1536,1552
day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
that what is done by the hook need
be updated.
Thanks for your point, have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 8: explain analyze is your friend
be used as a template, not just template1.
Moreover, template1 is a still usable database, that can be set with
whatever you want in new created, so it is not that special...
Well, that's grain for thoughts anyway;-)
Thanks, have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED
creation events.
I'm rather inclined to try the former option;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so
by the owner with a CREATE SCHEMA...
The later is simple and makes sense anyway for a newly created database.
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http
.
The alternative is to do it in C within the backend, but I would have
liked the plain SQL better. Just a mater of taste, I guess.
Pg backend philosophy: why writing SQL if you can do it in C? ;-)
I'll have a look at it if I have time, maybe over the week-end.
Thanks for your insight.
--
Fabien
foresee all consequences, but the current
situation is really inconsistent.
Any comment?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's
and _SOME as aggregate names?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
]: [35-1] LOG: aborting startup due to startup
process failure
So it is not a clean coredump, if some may be;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
to be particularly mnemonic for this
usage anyway.
Yep.
The standard EVERY is fine for postgres, the issue is only with
ANY/SOME.
Do you think that bool_and should be proposed anyway for homogeneity
with bool_or?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
Dear committers, dear hackers,
Subject: Re: [COMMITTERS] pgsql-server/src backend/utils/adt/acl.c ...
Ergo, my recommendation is to revert this change altogether. Fabien
should figure out the high-level description of what he wants to know
Please find attached as somehow requested
the public schema and maybe
some other stuff to be carefully selected, without bluring that important
distinction?
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
for your comments,
--
Fabien COELHO _ http://www.cri.ensmp.fr/~coelho _ [EMAIL PROTECTED]
CRI-ENSMP, 35, rue Saint-Honoré, 77305 Fontainebleau cedex, France
phone: (+33|0) 1 64 69 {voice: 48 52, fax: 47 09, standard: 47 08}
All opinions expressed here are mine
fully the rights of its database wrt
to other users without needing any super user privilege.
Good for students;-)
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
as
to help extension modules to be compiled, installed and even configured
easilly.
Is there any opposition to this move? Silence will mean consent;-)
Thanks in advance, have a nice day,
--
Fabien Coelho - [EMAIL PROTECTED]
---(end of broadcast
1 - 100 of 1578 matches
Mail list logo