??contrib

For information about 'contrib' see:

http://www.postgresql.org/docs/current/static/contrib.html

replace 'current' with 8.2

in the URL

Thanks

fts is working in my local server :)

hmmm, the 8.2 manual is missin ghte contributed section

Query: http://pastebin.com/S68BcAWa

I am getting this error: column "'" does not exist

LINE 11: AND 'Ramesh'::tsquery@@"'" & e.collecteur || " " || n...

iswaria: you're using identifier quotes instead of string quotes

single quotes -> strings; double-quotes -> identifiers

If i replace 8.2 for current, i am getting "The requested URL /docs/8.2/static/contrib.html was not found on this server."

so ITYW: '''' & e.collecteur || ' ' || nom_entier || ' ' || nom_complet & ''''

http://xkcd.com/718/ <-- lol!

still not working

iswaria, if you're starting a new project, do not use such an old version of pg

ok

heh. fermi's lack of a paradox.

endpoint_david: DRBD is O/S-level.

davidfetter: http://xkcd.com/715/

o/` i've got 99 problems o/`

(Action) puts on some danger mouse

and a bottle of beer ain't one?

99 might be :)

lol@ endpoint_david

Hint: No operator matches the given name and argument type(s). You might need to add explicit type casts.

what the meaning of it?

whatis "operator"?

whatis "argument"?

whatis "type"?

newbie2: show us your query?

??paste

For information about 'paste' see:

select * from c_order where c_order_id like '8%'

newbie2: use a paste site

ok

newbie2: hmm, you can't use like on an int/bigint.. you'd probably want c_order_id::text like '8%'

perhaps "like" is the operator

http://pgsql.privatepaste.com/a585b50f44

newbie2: have you found what those three words mean yet? here's a phrase you should look up too: "type cast" and "explicit type cast"

once you know what those mean, all the other words are plain english

newbie2 what are you trying to do?

newbie2 I'm guessing you want to find order_id number 8 ?

yes

that would be where = 8;

LIKE is for searching text

the only other thing I could imagine that you might be trying to do is find all orders that begin with the number 8

http://pgsql.privatepaste.com/a4d09ee453

http://pgsql.privatepaste.com/9497880b7f < that should work

??select

For information about 'select' see:

http://www.postgresql.org/docs/current/static/sql-selectinto.html :: http://www.postgresql.org/docs/current/static/sql-select.html#SQL-FOR-UPDATE-SHARE">http://www.postgresql.org/docs/current/static/sql-select.html#SQL-FOR-UPDATE-SHARE :: http://www.postgresql.org/docs/current/static/sql-select.html

http://www.postgresql.org/docs/current/static/plpgsql-statements.html :: http://www.postgresql.org/docs/current/static/plpgsql-statements.html#PLPGSQL-SELECT-INTO

http://pgsql.privatepaste.com/e7de6d4662 --> it work

thx all

Hi, im using 8.2.6. Is redirect_stderr configuration parameter enabled by default ? I can see that it's commented in the postgresql.conf. The postgresql doc doesnt say about the default value when the paramter is commented.

it's off by default

if you haven't edited the value, the commented-out value is the default one

and you can always show the actual value in effect, if the server is up, by connecting and using SHOW

RhodiumToad: that's convinient .. thanks for the info ..

the doc also doesnt say whether the directory specified in the log_directory parameter will be automatically created when it's unavailable .. but i guess i can find out by trying :)

iirc it won't be

ok ..

and if i want to see who issued what sql statements, is the log level of 'notice' sufficient ?

see log_statement

(or log_min_duration_statement=0)

sorry .. just reached that section in the docs .. :)

make sure you set log_line_prefix to something suitably comprehensive

i used '<%t %u@%e> '

time to restart to test things out

thanks alot for the help

albert_kam: you really need %p

why PID ?

otherwise you can't sort out which session is which

ohh ..

ok2 .. adding

hello

Hi, i wonder why my log outputs so much from a single select statement ? the log file for a single sql select : http://pastie.org/884285 .. the conf file for the logging : http://pastie.org/884290

i dont expect to see the parse or bind duration, and i dont want to see other statements issued on pg_catalog n such. i prefer only seeing the info for the issued sql statement ..

a simpler output that's more readable

what did i miss in the configuration ? :)

albert_kam: how are you executing those queries?

merlin83: i did ctrl+e (execute sql) on aquadata studio app for the sql "select * from mytable"

lemme try it on psql

albert_kam: you'll see parse/bind/execute for queries executed using PQexecParams

albert_kam: the three phases have to be logged separately because for prepared queries, there might be one parse and multiple binds, and with some client interfaces there can be multiple executes per bind

wow, the log output from psql is very clean and is what i wanted, lol

(Action) wonders where WIT is

albert_kam: looks like "aquadata studio" is doing a lot of its own metadata queries

very inefficiently, too

I mean, a whole roundtrip to the server _for each column_ just to check whether it's declared not null?

I see ..

on a remote server that'd be seriously painful

thats the gui tool we use here as the sql manager, to do sql devel .. it's a shame it issued many sql statements ..

Strange why it's WIT, should be WIB. Btw, im in Jakarta, Indonesia. But perhaps people dont usually know Indonesia, but they know Bali :)

yeah, I guessed that from running select * from pg_timezone_names where abbrev='WIT';

morning

albert_kam: selamat pagi!

albert_kam: oh, well, for me at least :)

kll: selamat siang here :)

RhodiumToad: your opinion on the database client now makes me to find other alternatives .. any recommendation for all-purpose GUI database ? perhaps http://torasql.com/ ? :)

??pgpool

For information about 'pgpool' see:

http://pgpool.projects.postgresql.org/pgpool-HA/en/

http://pgpool.projects.postgresql.org/

http://pgfoundry.org/projects/pgpool/

(Action) isn't the person to ask about GUIs

RhodiumToad: my apology, hee hee

"xterm -e psql" :-)

i have problem with pg_tablespace_size('mytable') and i have error does not exist

but select * from mytable work fine ?

gintek; pg_total_relation_size('mytable')

any idea what i`m doing wrong with this stupid problem ?

gintek; tablespaces are something completely different

johto: why in manual pg_database_size(name) Disk space used by the database with the specified name

that would work if you want to know the size of the database, but I don't see how that's related

pg_total_relation_size work fine

johto: thx

databases, tablespaces and tables (relations) are completely different things

?? tablespace

For information about 'tablespace' see:

http://people.planetpostgresql.org/xzilla/index.php?/archives/322-tablespace-management-variables.html

quick question concerning pgadmin

are there any Windows builds available for releases since 1.10.2?

??mysql

For information about 'mysql' see:

http://sql-info.de/mysql/gotchas.html :: http://www.raditha.com/mysql/mysql2pgsql.php :: http://www.in-nomine.org/~asmodai/mysql-to-pgsql.html :: http://pgfoundry.org/projects/mysql2pgsql/

http://search.cpan.org/search?query=SQL%3A%3ATranslator&mode=all :: http://www.metatrontech.com/wpapers/mysql2postgresql.pdf :: http://www.data-conversions.net/products.php?prod_num=5&&dest=MENU&&ID=200 :: http://pgfoundry.org/projects/mysqlcompat

Hi

Have you already an webservice using SOAP and postgres?

hi, can someone explain the window functions 'lead' and 'lag'?

hello, is it possible to view last changes made in some database if they are not executed through psql? something like history?

pif; "lead" is the row following the current row and "lag" is the previous row

ok

hi all

there was a way to set up the psql to show the row with the null values in order to distinguish from a empty row

but I don't remember which

\pset null foo

thanks

RhodiumToad: ?

?

you use FreeBSD, no ?

for you is it normal that /usr/local/lib/libpgport.a isn't provided by the -client package ?

I filled a PR for mapserver which require this file (http://www.freebsd.org/cgi/query-pr.cgi?pr=144985) but the fix of the package maintener is only to add a dependency on -server ...

it looks strange to me that you cannot build mapserver with postgis support without building the -server port

what is it referencing libpgport for?

it's required by mapserver when you enable PostGIS support

what functions is it trying to use?

I don't known .. the port fails to build without it

with :

/usr/bin/ld: cannot find -lpgport

gmake: *** [shp2img] Error 1

hm

but I don't understand why -lpgport is listed with pg_config --libs and that that file (/usr/local/lib/libpgport.a) is missing

(on a machine where only -client is installed)

the client/server division in the freebsd port of pg is a bit questionable in parts, though that's partly due to the fact that the project itself doesn't provide any clear distinction

do you think I should fill a PR for this ?

probably

ok thanks

pg_config is a problem in itself because of confusion over whether it reflects the server build or the client

it really needs splitting into two

yep ..

there should be a server version that reflects what is needed to build server modules, and a client version

I'll include that in the PR

since there are clients like dbd::pg and (I assume) mapserver that want to use it for client builds, packaging systems tend to put in a version suited to the client needs,

which then buggers up any attempt to build server modules correctly

there should really be a libpq port in FreeBSD

hey everyone. When I have a pl/pgsql function with "rows:=SELECT colname FROM tablename WHERE colname=val;", how can I test if that query returned any results?

IF FOUND THEN .

johto, cool, thanks

that statement is wrong anyway

you want SELECT INTO

(and FOUND isn't set by assignment statements)

RhodiumToad, yeah. I just figured that out :) Thanks. I just do SELECT colname FROM tablename WHERE colname=var;IF NOT FOUND THEN blah....

no

without the assignment

that'll fail too

ok

are you trying to get the actual value of "colname" or just check whether the row exists?

just checking if it exists

I don't need the value

if exists(select * from tablename WHERE colname=var) then ...

ah, that is better. Thanks

hey guys

whats the cost of a simple query?

like, is there a medium or large startup cost just to execute the query?

Diablo-D3: there's some overhead of parsing and planning the query

you can avoid both by using prepared statements

okay so like

lets say I just want to return rows, and do no real thinking outside of just simple column matching

I have 1 server, a million rows

you just want to get all the rows from the server?

vs

100 servers, the data set split up

how much would I be losing if I executed the same simple query on every one