RE: JDBC Meta-data performance for Foreign Keys - Mailing list pgsql-jdbc

From ldh@laurent-hasson.com
Subject RE: JDBC Meta-data performance for Foreign Keys
Date
Msg-id MN2PR15MB2560BA59E0E4FD739B388BFD85AE9@MN2PR15MB2560.namprd15.prod.outlook.com
Whole thread Raw
In response to Re: JDBC Meta-data performance for Foreign Keys  (Mark Rotteveel <mark@lawinegevaar.nl>)
List pgsql-jdbc

   >  -----Original Message-----
   >  From: Mark Rotteveel <mark@lawinegevaar.nl>
   >  Sent: Saturday, October 2, 2021 04:57
   >  To: pgsql-jdbc@lists.postgresql.org
   >  Subject: Re: JDBC Meta-data performance for Foreign Keys
   >  
   >  On 02-10-2021 08:29, ldh@laurent-hasson.com wrote:
   >  > Lots of code refactoring on our end, but it was worth it as the results
   >  are quite good with a ~70% improvement!
   >  >
   >  > The only gotcha is what I tried to also optimize the calls for indices, I
   >  was surprised that the jdbc method DatabaseMetaData.getIndexInfo()
   >  doesn't take a table pattern like other "getXXXInfo" methods. For now
   >  we have decent performance, so dropping to a custom query to optimize
   >  indices is something we can do later.
   >  >
   >  >
   >  >    pre-optimization
   >  >    ==========================================
   >  >    [..]
   >  >    Views: 362 in 22.00ms or 0.1%
   >  >    ------------------------------------------
   >  >    Total: 19,667.00ms
   >  >
   >  >
   >  >    post-optimization
   >  >    ==========================================
   >  >    [..]
   >  >    Views: 362 in 113.00ms or 1.9%
   >  >    ----------------------------------------------------------------------------
   >  >    Total: 5,809.00ms
   >  
   >  I notice that you seem to be requesting views separately. You should be
   >  able to do that in a single call to DatabaseMetaData.getTables together
   >  with retrieving the tables.
   >  
   >  Mark
   >  --
   >  Mark Rotteveel
   >  


It's the way we do the logging for processing... But in any case, we are talking a few milliseconds here so not
somethingI was paying attention to specifically.
 

Thank you,
Laurent.

pgsql-jdbc by date:

Previous
From: Thomas Kellerer
Date:
Subject: Can't run CREATE PROCEDURE with new Postgres' 14 new SQL-standard body
Next
From: Dave Cramer
Date:
Subject: [pgjdbc/pgjdbc] 3cf175: fix: Regressions caused by PR#1949 (#2266)