THE SQL Server Blog Spot on the Web

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


You searched for the word(s):
Showing page 1 of 48 (477 total posts) < 1 second(s)
  • re: Bushy Joins

    thanks Rob. I should add that prior to this I had never written a RIGHT JOIN. I have always placed the outer join table with LEFT JOIN. It makes no difference in SQL, but if we are forcing the join order, then we would use a right join in certain situations.
    Posted to Joe Chang (Weblog) by jchang on April 9, 2014
  • re: Bushy Joins

    I am going to SQL Sat Atlanta. I also put in for Redmond, Oslo and Paris. Three times I have been involved with Chicago ISVs, all of them are SQL disasters. What's sad is that these could be have fixed had anyone had a basic knowledge of SQL Server. Anyways, what I posted here is the dangerous thing, as you said in your session. But I am curious ...
    Posted to Joe Chang (Weblog) by jchang on April 2, 2014
  • Bushy Joins

    A great session by Adam Machanic at SQL Saturday Boston the previous weekend on methods to influence the query optimizer while still letting it do its task. The gist of this is that while SQL Server has what are called Query Hints, there are adverse consequences. The Join Hints (Loop, Hash and Merge) &quot;specify that the query optimizer ...
    Posted to Joe Chang (Weblog) by jchang on April 2, 2014
  • Build your own server with Supermicro motherboards

    I used to build white box servers because there were usually enough spare parts left over from upgrade projects. (management did not see the need for non-production systems, so I arranged for there to be spare parts). But since 2002 or so, I have been buying Dell PowerEdge servers for my own test environment. This was in part because of the ...
    Posted to Joe Chang (Weblog) by jchang on April 2, 2014
  • re: IO Queue Depth Strategy

    for old posts like this, please email or I may not find your comment. I was not aware of significant IO strategy changes from 2008R2 to 2012, but there was a significant somewhere in the 2008 RTM-R2 period. Assuming that there were no hardware changes, FC HBA queue depth setting etc. If I had to guess, the only that comes to mind that would ...
    Posted to Joe Chang (Weblog) by jchang on April 1, 2014
  • re: Hekaton and Benchmarks?

    Excellent, I had not realized there was a limit for memory optimized tables. Is the limit also dependent on system memory? Also, I failed to consider that the 2 x 12 core 2590 tps-E result has a db size of 10.7TB, way bigger than memory. So 10X would mean a 100TB database. There is a rigid requirement in the TPC benchmarks for database size to ...
    Posted to Joe Chang (Weblog) by jchang on March 14, 2014
  • Hekaton and Benchmarks?

    It is rather curious that the two TPC-E benchmarks results published for SQL Server 2014 did not employ the memory optimized tables and natively compiled procedures, given that Hekaton is the hallmark feature of 2014. Potential for 30X gain in transaction processing have been cited in presentations and whitepapers. No explanation has ...
    Posted to Joe Chang (Weblog) by jchang on March 13, 2014
  • re: Column Store, Parallelism and Decimal

    Ok, you are right, the difference is very likely to be due to Row versus Batch mode. On SQL 2014 CTP2, tested on a clean install system, the column store in Row mode has CPU very close to the normal page/row table. &nbsp;I only did a quick check for SQL 2012 sp1 on my laptop, I did not have memory for the full (page/row) table scan. And the 2012 ...
    Posted to Joe Chang (Weblog) by jchang on March 4, 2014
  • re: Column Store, Parallelism and Decimal

    thanks, I had been looking for that link. But the SQL Server 2012 behavior is correct. It is the 2014 CTP2 has seems to be stupid. Could it be that Microsoft documented this effect in 2012, but did not activate it until 2014? In any case, this is similar to the Bitmap filter, which is only activated in parallel plans, even though it could help in ...
    Posted to Joe Chang (Weblog) by jchang on March 3, 2014
  • re: Fake statistics, and how to get rid of them

    Thomas Kesjer and I discussed statistics workarounds a few years back. I recall that he noted that when the true statistics (in stats_stream, formerly statblob) were sampled at 100%, the ROWCOUNT and PAGEGOUNT fake values were ignored, but not if sampling were less than 100%. TK might comment on this, or you could look into it as my memory is not ...
1 2 3 4 5 Next > ... Last »
Powered by Community Server (Commercial Edition), by Telligent Systems
  Privacy Statement