Unsupported Screen Size: The viewport size is too small for the theme to render properly.

Bug Function Names Are Not Included In The Autocomplete

forums forums SQLyog SQLyog: Bugs / Feature Requests Bug Function Names Are Not Included In The Autocomplete

  • This topic is empty.
Viewing 8 reply threads
  • Author
    Posts
    • #12426
      Pentium10
      Member

      when I have tried to reference a function name in the select fragment it was not included in the list of choices

      Please fix ASAP.

      http://screencast.com/t/3DzuigdD

    • #32552
      peterlaursen
      Participant

      I have cleaned up here. I misunderstood.

      You are right. The function *name* of a stored fucntion should display in acutocomplete- I wonder it it is because there is text on the right side.

    • #32553
      vishal.pr
      Member

      @Pentium10,

      Can you rebuild the tag files and try. To rebuild tags, in main menu goto Powertools->Rebuild tags.

      Regards,

      Vishal P.R

    • #32554
      Pentium10
      Member
      'vishal.pr' wrote:


      @Pentium10
      ,

      Can you rebuild the tag files and try. To rebuild tags, in main menu goto Powertools->Rebuild tags.

      Regards,

      Vishal P.R

      That says there Stop Rebuilding Tags for a while. I have a huge system with over 1030 databases.

      I think you should improve the order of that thread a bit.

      So it should queue in front, with higher priority the databases that are expanded/use(d)/favorited in the object browser first.

      Anyway does that some kind of local caching?

      Or whenever I open a new connection it tries to build the tags? That will take hours on a large database system like I have, so I can't use SQLYog for hours until the tags are not finished. That is not acceptable, and for example our admin users will ban me because I do a lot of load on the system by having so many connections open for hours and doing heavy stuff like schema readings.

      Further more I see a lot of processes after I start SQLYog that are very slow: SHOW CREATE TABLE …….., and all these are making a huge load on the database because for each column MySQL checks permissions, so runs a query on: information_schema`.`COLUMNS`

      And the order there is in alphabetic order, and there is no priority on the parsed databases, which is totally WRONG, as it will take down my database.

      See the heavy load on the server and the time the connection are kept open (10 minutes and only got parsed until databases with letter F)

      http://screencast.com/t/yhrIXaBGH

      It queries tables that are never used in the GUI.

      Update: The build tags did not finished, it timed out after 1 hour. And only half of the database tables were processed.

      Update: I've tried this on a server where only a few database are, and the stored procedures/functions are not tagged, and rebuild tags does not help, and they are not included in the hint.

    • #32555
      peterlaursen
      Participant

      1) The SQLite databases use by autocomplete are built on a per-connection basis. Do you need access to all databases from same connection? Note that in the SQLyog 'connection manager' you can specify specific databases only. That will result in much smaller local databases and

      2) Can you find the 'Tags' folder in your user profilesqlyog folder. How big is the biggest file in here?

      3) I don't understand why it should stop/time out. Do you use (HTTP- or SSH-)tunnel?

      4) We are just executing simple SHOW statements (SHOW DATABASES, SHOW TABLES LIKE 'base_table' etc.) to get the meta information inserted to lcoal database. We cannot reduce the load from that. And querying I_S will cause similar (or higher on some mySQL versions) load. I wonder how well you server is configured? This article is probably the best entry point http://www.mysqlperformanceblog.com/2006/09/29/what-to-tune-in-mysql-server-after-installation/ (though maybe with MySQL 5.5 it may not be quite up to date).

    • #32556
      Pentium10
      Member
      'peterlaursen' wrote:

      1) The SQLite databases use by autocomplete are built on a per-connection basis. Do you need access to all databases from same connection? Note that in the SQLyog 'connection manager' you can specify specific databases only. That will result in much smaller local databases and

      2) Can you find the 'Tags' folder in your user profilesqlyog folder. How big is the biggest file in here?

      3) I don't understand why it should stop/time out. Do you use (HTTP- or SSH-)tunnel?

      4) We are just executing simple SHOW statements (SHOW DATABASES, SHOW TABLES LIKE 'base_table' etc.) to get the meta information inserted to lcoal database. We cannot reduce the load from that. And querying I_S will cause similar (or higher on some mySQL versions) load. I wonder how well you server is configured? This article is probably the best entry point http://www.mysqlperformanceblog.com/2006/09/29/what-to-tune-in-mysql-server-after-installation/ (though maybe with MySQL 5.5 it may not be quite up to date).

      1. I don't need access to all databases, probably need access to 50, I will try to list the number of databases I need. I will post back the findings.

      2. One file is very big: 100 MB, more precisely 104,878,080 bytes others are at 1MB

      3. I am connecting on a gigabit network inside the building (no tunneling)

      4. Server is configured well as it handles so much trafic on all those database, it's a super machine. When I stop SQLYog the load goes down below 0.8

      How much work would be to priorities parsing expanded/use(d)/favorited objects first?

      Update: I've tried listing a couple of databases in the connection wizard, and now it is much faster and works better, and stored functions names are used in autohint. Thanks for this tip. Could I request to put this in your FAQ.

      I got this working now, but there are other questions in this thread, so please keep discussing on that and we should not close for now.

    • #32557
      peterlaursen
      Participant

      The idea to first process the current database context is a good request really. Our developer team will check how easy it can be done.

    • #32558
      Pentium10
      Member
      'peterlaursen' wrote:

      The idea to first process the current database context is a good request really. Our developer team will check how easy it can be done.

      Can you post this request also in the Google Issues?

    • #32559
      peterlaursen
      Participant

      I think you did yourself and I have set status as 'accepted'.

      http://code.google.c…/detail?id=1687

Viewing 8 reply threads
  • You must be logged in to reply to this topic.