|
|
You're in the
sql server data warehouse
group:clustered index or simple constraint in a fact table?
sql server data warehouse:
What do you recommand to insure the integrity of a fact table? a cluster index? or a constraint? or an SQL script? do you recommand to allways having a clustering index on a fact table? or just on the dimension tables and 1 index by linked column in the fact table? I don't want to test each possibility, so I want your help ;) Jerome. --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28
What do you mean by integrity? A clustered index does not have to be unique.
There are very few times I will have any table without a clustered index, DW or OLTP. Linked column? Do you mean a foreign key? -- Ray Higdon MCSE, MCDBA, CCNA --- [quoted text, click to view] "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message news:O06B%23Zj$DHA.3184@TK2MSFTNGP09.phx.gbl... > Hi, > > What do you recommand to insure the integrity of a fact table? > a cluster index? > or a constraint? > or an SQL script? > > do you recommand to allways having a clustering index on a fact table? or > just on the dimension tables and 1 index by linked column in the fact table? > > I don't want to test each possibility, so I want your help ;) > > Jerome. > > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > >
Also, these links may help
http://www.microsoft.com/sql/techinfo/administration/2000/security/default.asp http://www.sql-server-performance.com/olap_performance.asp -- Ray Higdon MCSE, MCDBA, CCNA --- [quoted text, click to view] "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message news:O06B%23Zj$DHA.3184@TK2MSFTNGP09.phx.gbl... > Hi, > > What do you recommand to insure the integrity of a fact table? > a cluster index? > or a constraint? > or an SQL script? > > do you recommand to allways having a clustering index on a fact table? or > just on the dimension tables and 1 index by linked column in the fact table? > > I don't want to test each possibility, so I want your help ;) > > Jerome. > > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > >
Yes, sorry, I talk about a unique clustered index.
Actually my fact table has a primary key which used all the foreign key columns. I've set all my indexes in a separate file to improove the performance, but because Most part of the time, SQL Server used the cluster index instead-of using the other indexes, the indexes files are not used, so no performance improovment.. If I drop the cluster index and keep only the indexes associated with each foreign key cloumn, the queries are not rightly optimized and SQL don't used some indexes. Fact table - ForeKey1 --> idx 1 - ForeKey2 --> idx 2 - ForeKey3 --> idx 3 - ForeKey4 --> idx 4 - Measure 1 - Measure 2 My usage is not an OLAP cube, but only a reporting tool. Well, I'm disapointed when I see how SQL Server optimize the queries. Also the query plan never represent the real value of the result produce by the query but my statistics are created and updated. So I work to optimize all of this. Thanks for your guide. "Ray Higdon" <sqlhigdon@nospam.yahoo.com> a écrit dans le message de news:%233W1VNl$DHA.2632@TK2MSFTNGP12.phx.gbl... [quoted text, click to view] > What do you mean by integrity? A clustered index does not have to be unique. > There are very few times I will have any table without a clustered index, DW > or OLTP. > > Linked column? Do you mean a foreign key? > > -- > Ray Higdon MCSE, MCDBA, CCNA > --- > "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message > news:O06B%23Zj$DHA.3184@TK2MSFTNGP09.phx.gbl... > > Hi, > > > > What do you recommand to insure the integrity of a fact table? > > a cluster index? > > or a constraint? > > or an SQL script? > > > > do you recommand to allways having a clustering index on a fact table? or > > just on the dimension tables and 1 index by linked column in the fact > table? > > > > I don't want to test each possibility, so I want your help ;) > > > > Jerome. > > > > > > > > --- > > Outgoing mail is certified Virus Free. > > Checked by AVG anti-virus system (http://www.grisoft.com). > > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > > > > > > --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28
It depends on your queries as to what will be used. Be aware that clustered
indexes sit at the leaf level of the data so you can not split them up. How are you determining that the queries are not optimized? The best way I've found for this is to use the "set statistics IO on" command and look at logical IO, which is the number of times a page is hit (not number of pages). You might post your queries and we can look at why optimizer may choose one over the other. -- Ray Higdon MCSE, MCDBA, CCNA --- [quoted text, click to view] "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message news:OhQeYss$DHA.2432@TK2MSFTNGP11.phx.gbl... > Yes, sorry, I talk about a unique clustered index. > Actually my fact table has a primary key which used all the foreign key > columns. > > I've set all my indexes in a separate file to improove the performance, but > because Most part of the time, SQL Server used the cluster index instead-of > using the other indexes, the indexes files are not used, so no performance > improovment.. > > If I drop the cluster index and keep only the indexes associated with each > foreign key cloumn, the queries are not rightly optimized and SQL don't used > some indexes. > > Fact table > - ForeKey1 --> idx 1 > - ForeKey2 --> idx 2 > - ForeKey3 --> idx 3 > - ForeKey4 --> idx 4 > - Measure 1 > - Measure 2 > > My usage is not an OLAP cube, but only a reporting tool. > > Well, I'm disapointed when I see how SQL Server optimize the queries. Also > the query plan never represent the real value of the result produce by the > query but my statistics are created and updated. > > So I work to optimize all of this. > > Thanks for your guide. > > "Ray Higdon" <sqlhigdon@nospam.yahoo.com> a écrit dans le message de > news:%233W1VNl$DHA.2632@TK2MSFTNGP12.phx.gbl... > > What do you mean by integrity? A clustered index does not have to be > unique. > > There are very few times I will have any table without a clustered index, > DW > > or OLTP. > > > > Linked column? Do you mean a foreign key? > > > > -- > > Ray Higdon MCSE, MCDBA, CCNA > > --- > > "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message > > news:O06B%23Zj$DHA.3184@TK2MSFTNGP09.phx.gbl... > > > Hi, > > > > > > What do you recommand to insure the integrity of a fact table? > > > a cluster index? > > > or a constraint? > > > or an SQL script? > > > > > > do you recommand to allways having a clustering index on a fact table? > or > > > just on the dimension tables and 1 index by linked column in the fact > > table? > > > > > > I don't want to test each possibility, so I want your help ;) > > > > > > Jerome. > > > > > > > > > > > > --- > > > Outgoing mail is certified Virus Free. > > > Checked by AVG anti-virus system (http://www.grisoft.com). > > > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > > > > > > > > > > > > > --- > Outgoing mail is certified Virus Free. > Checked by AVG anti-virus system (http://www.grisoft.com). > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > >
there is too many queries
because the users can creates the queries they want on the fly. there is some predefined reports, and I'll optimize these specific reports so a user can filter on 1 or more columns, links 1 or more dimension tables etc... Also I've a lot of distinct count measures I'll check the IO stats this week "Ray Higdon" <sqlhigdon@nospam.yahoo.com> a écrit dans le message de news:eDtNUXt$DHA.2216@TK2MSFTNGP10.phx.gbl... [quoted text, click to view] > It depends on your queries as to what will be used. Be aware that clustered > indexes sit at the leaf level of the data so you can not split them up. How > are you determining that the queries are not optimized? The best way I've > found for this is to use the "set statistics IO on" command and look at > logical IO, which is the number of times a page is hit (not number of > pages). You might post your queries and we can look at why optimizer may > choose one over the other. > > -- > Ray Higdon MCSE, MCDBA, CCNA > --- > "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message > news:OhQeYss$DHA.2432@TK2MSFTNGP11.phx.gbl... > > Yes, sorry, I talk about a unique clustered index. > > Actually my fact table has a primary key which used all the foreign key > > columns. > > > > I've set all my indexes in a separate file to improove the performance, > but > > because Most part of the time, SQL Server used the cluster index > instead-of > > using the other indexes, the indexes files are not used, so no performance > > improovment.. > > > > If I drop the cluster index and keep only the indexes associated with each > > foreign key cloumn, the queries are not rightly optimized and SQL don't > used > > some indexes. > > > > Fact table > > - ForeKey1 --> idx 1 > > - ForeKey2 --> idx 2 > > - ForeKey3 --> idx 3 > > - ForeKey4 --> idx 4 > > - Measure 1 > > - Measure 2 > > > > My usage is not an OLAP cube, but only a reporting tool. > > > > Well, I'm disapointed when I see how SQL Server optimize the queries. Also > > the query plan never represent the real value of the result produce by the > > query but my statistics are created and updated. > > > > So I work to optimize all of this. > > > > Thanks for your guide. > > > > "Ray Higdon" <sqlhigdon@nospam.yahoo.com> a écrit dans le message de > > news:%233W1VNl$DHA.2632@TK2MSFTNGP12.phx.gbl... > > > What do you mean by integrity? A clustered index does not have to be > > unique. > > > There are very few times I will have any table without a clustered > index, > > DW > > > or OLTP. > > > > > > Linked column? Do you mean a foreign key? > > > > > > -- > > > Ray Higdon MCSE, MCDBA, CCNA > > > --- > > > "Jéjé" <willgart@_A_hotAmail_A_.com> wrote in message > > > news:O06B%23Zj$DHA.3184@TK2MSFTNGP09.phx.gbl... > > > > Hi, > > > > > > > > What do you recommand to insure the integrity of a fact table? > > > > a cluster index? > > > > or a constraint? > > > > or an SQL script? > > > > > > > > do you recommand to allways having a clustering index on a fact table? > > or > > > > just on the dimension tables and 1 index by linked column in the fact > > > table? > > > > > > > > I don't want to test each possibility, so I want your help ;) > > > > > > > > Jerome. > > > > > > > > > > > > > > > > --- > > > > Outgoing mail is certified Virus Free. > > > > Checked by AVG anti-virus system (http://www.grisoft.com). > > > > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > > > > > > > > > > > > > > > > > > > > --- > > Outgoing mail is certified Virus Free. > > Checked by AVG anti-virus system (http://www.grisoft.com). > > Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28 > > > > > > --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.600 / Virus Database: 381 - Release Date: 2004-02-28
Don't see what you're looking for? Try a search.
|
January 2003
March 2003
July 2003
August 2003
September 2003
October 2003
November 2003
December 2003
January 2004
February 2004
March 2004
April 2004
May 2004
June 2004
July 2004
August 2004
September 2004
October 2004
November 2004
December 2004
January 2005
February 2005
March 2005
April 2005
May 2005
June 2005
July 2005
August 2005
September 2005
October 2005
November 2005
December 2005
January 2006
February 2006
March 2006
April 2006
May 2006
June 2006
July 2006
August 2006
September 2006
October 2006
November 2006
December 2006
January 2007
February 2007
March 2007
April 2007
May 2007
June 2007
July 2007
August 2007
September 2007
October 2007
November 2007
| home · blog · groups | Questions? Comments? Contact the dn |