Got more questions? Find advice on: ASP | XML | Regular Expressions | Windows
in Search
Welcome to SqlAdvice Sign in | Join | Help

Browse by Tags

All Tags » Tips And Tricks
  • Full Text Indexing Overhead

    I think that there's a tendency among DBAs and SQL Server wonks to be a bit dismissive of Full Text Indexing. I think that in a lot of cases people tend to think of it in terms of adding significant overhead. For a long time I actually looked at it as being a more powerful, but more EXPENSIVE, form of searching. Then I actually had the chance to ...
    Posted to REPEATABLE READ (Weblog) by michael.k.campbell on November 9, 2007
  • Burned again by HyperThreading on SQL Server 2000

    One of my clients recently had a problem where a query with a semi-measly execution cost of .3 was taking 25 seconds to return. Worse, forcing the query to use a suitable index (instead of the PK with accompanying bookmarks) had a cost of 4.8 and returned instantly. Moreover, the query taking 25 seconds was only iterating over a ...
    Posted to REPEATABLE READ (Weblog) by michael.k.campbell on February 13, 2007
  • Micro Indexing

    It's almost a rule of thumb to avoid indexing strategies on tables with less than a few thousand rows. The reason, of course, is that SQL Server will frequently just ignore indexes on 'tiny' tables - preferring instead to perform table scans. However, in environments with heavy load, and with borderline memory bottlenecks (i.e. heavy memory ...
    Posted to REPEATABLE READ (Weblog) by michael.k.campbell on January 25, 2007
  • Using Virtualization to fight ... Virtualization

    VPN Software should come with big warning labels: ''WARNING: Does not play well with others.''As a consultant I frequently need to gain access to remote networks via VPN. This means I'm frequently hosing myself by installing yet another VPN client. That, and I hate connecting to the vast majority of VPNs because they hose my network settings ...
    Posted to REPEATABLE READ (Weblog) by michael.k.campbell on April 20, 2006
  • SQL Server Unit Testing: Sprocs

    As a consultant, it's not uncommon for me to need to rewrite a poorly performing query or sproc without a full understanding of the underlying biz rules or context. Improved performance is the stated goal, but data fidelity is implied as critical. Enter unit testing - something many of us rely upon heavily during 'normal' development at ...
    Posted to REPEATABLE READ (Weblog) by michael.k.campbell on April 6, 2006
  • Nibbling Deletes - Intro

    So, your application was a success. You're getting oodles and oodles of traffic. Only, it's building up faster than you had anticipated, and after a bit of calculation, you've determined that if the present trends persist, you'll have 24 TB of data within 1 year. Only, you really don't need data > 3 months old in the main OLTP ...
    Posted to REPEATABLE READ (Weblog) by RepeatableRead on September 20, 2005
  • Max Degree of Parallelism

    In a previous life (actually two previous lives), I was a production DBA. One of the environments that I inherited had some ... issues. There were a number of Clustered SQL Server installations -- oodles of hardware, all churning to keep product inventory, customer information, and catalog browsing going for a number of web farms for a ...
    Posted to REPEATABLE READ (Weblog) by RepeatableRead on May 24, 2005
  • Holy Mind Bending Bafflement Batman!!

    I'm still looking into this one myself. But at this point I'm thinking I may have stumbled on to a bug. Someone explain what's up with the following: -- returns 0123456789 SELECT ISNULL(NULL,'0123456789') -- returns NULL SELECT NULLIF('0123456789','0123456789') -- works, because 01234 is shorter than 0123456789 SELECT ...
    Posted to REPEATABLE READ (Weblog) by RepeatableRead on December 3, 2004
  • IDIOM: Using Dynamically created table names without resorting to Dynamic SQL

    I hate Dynamic SQL. It's as lame as in-line SQL. So I lose sleep when I'm 'forced' to use Dynamic-SQL to solve problems. Oh... I'll use it in an absolute pinch, but after the 'pinch' is over I try flipping the problem on its head over and over to see if I can come up with a non-cheesy solution. That being said, a common place many people ...
    Posted to REPEATABLE READ (Weblog) by RepeatableRead on July 27, 2004
  • Custom Templates

    T-SQL is vast. Think about it: it allows for the creation and control of everything from indexes, tables, data-cubes, full-text indexes, full/differential/log backups, statistic optimization, and on and on and on. Unless you've been using SQL Server since version 4.2 there's entirely too much syntax (along with nuances, variations, etc) to master. ...
    Posted to REPEATABLE READ (Weblog) by RepeatableRead on June 7, 2004