THE SQL Server Blog Spot on the Web

Welcome to SQLblog.com - The SQL Server blog spot on the web Sign in | |

Search

You searched for the word(s):
Showing page 1 of 14 (133 total posts) < 1 second(s)
  • re: Choosing SEQUENCE instead of IDENTITY? Watch your step.

    I guess, several hours of your life spared much more hours to other people :-). Anyway, if it generates memory dump - it is definitely bug. No _SQL_ error should generate dump in a normally functioning system. I think, you should open a Connect item.
    Posted to Maria Zakourdaev (Weblog) by Michael Zilberstein on September 8, 2014
  • re: For want of a nail

    Len, Microsoft told us that 2 other users reported same bug in the same SQL Server version and that nobody experienced anything like this in other versions. So since they couldn't find the root cause and since it seems that in other version bug doesn't exist, they recommended us to upgrade to next service pack. We did an upgrade and indeed the ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on August 21, 2014
  • re: Reading temporary table from another session

    There're plenty of solutions to monitor progress: use normal and not temporary table, use global temp table as you've proposed, print some outputs (I prefer to use RAISERROR with severity 10) etc. The problem was - I've thought about monitoring progress too late, so had to find another solution. No doubt, if only I could think about it in advance, ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on July 14, 2014
  • re: Reading temporary table from another session

    Chris, I'm not sure that's possible using SQL Server tools. DMVs usually don't pay much attention to NOLOCK hint. If they're locked - they're locked. Logically, those DMVs sit on top of system pages like PFS, GAM and SGAM. When you add rows (and probably allocate new pages to your table), it is logical that while in transaction, your bits in PFS ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on March 19, 2014
  • re: Reading temporary table from another session

    Chris, I'm not sure that's possible using SQL Server tools. DMVs usually don't pay much attention to NOLOCK hint. If they're locked - they're locked. Logically, those DMVs sit on top of system pages like PFS, GAM and SGAM. When you add rows (and probably allocate new pages to your table), it is logical that while in transaction, your bits in PFS ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on March 19, 2014
  • re: Grouping events into intervals using hedgehog rolling style

    Maria, interesting and very useful - I have almost same task at one of my databases. Several questions: 1) For 2nd step did you try self join instead of LAG? I once was very surprised to find out that old solution can be more efficient. http://sqlblog.com/blogs/michael_zilberstein/archive/2012/03/14/42332.aspx 2) Entire task is very similar to ...
    Posted to Maria Zakourdaev (Weblog) by Michael Zilberstein on December 3, 2013
  • re: Updateable columnstore index gotchas

    Thanks! I see that you've addressed all the issues in great detail. Glad to be an inspiration, even in such circumstances :-). At least now I can say that I have basic understanding of the flow - had no idea that bulk insert of 100K+ rows can bypass deltastore and go straight to the index. Same about dictionary-related information. And that's ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on December 2, 2013
  • re: Updateable columnstore index gotchas

    tobi, thanks! That's what I needed. BOL says: -- Reorganizing is not required for moving CLOSED rowgroups into the columnstore. The tuple-mover process will eventually find all CLOSED rowgroups and move them. However, the tuple-mover is single-threaded and might not move rowgroups fast enough for your workload. That was my point all along - why ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on December 2, 2013
  • re: Updateable columnstore index gotchas

    tobi, it sounds strange - data in deltastore is kept in a usual rowstore format. Segments should be closed first and then compressed - only then data is in the index - automatically. So I don't see how you can just copy from deltastore to columnstore - those are 2 different formats.
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on December 2, 2013
  • re: Updateable columnstore index gotchas

    Remus, actually I see the answer to my question in your great article (read it in August - enough time to forget :-) http://rusanu.com/2013/06/11/sql-server-clustered-columnstore-indexes-at-teched-2013/ -- When handling deltastores the columnar storage advantages are dimmed. Deltastores are row mode storage so the entire row has to be read, not ...
    Posted to Michael Zilberstein (Weblog) by Michael Zilberstein on December 1, 2013
1 2 3 4 5 Next > ... Last »
Powered by Community Server (Commercial Edition), by Telligent Systems
  Privacy Statement