Solved

optimize select distinct

Posted on 2010-09-23
7
525 Views
Last Modified: 2012-05-10
Hello,

How can I optimize this select distinct :
Select      Distinct      
      TEST.SOC_CODE CODE_SOCIETE,
      TEST.CLI_CODE CODE_CLIENT,
      TEST.TEST_ORIGINE_CLIENT CODE_ORIGINE_CLIENT,
      TEST.AGE_CODE CODE_AGENCE,
      TEST.TEST_CODE NUMERO_TESTIS,
      TEST.TEST_VERSION CODE_VERSION_TESTIS,
      case when TEST.TEST_NATURE_LIVRAISON = 'L' then 'LIVRE' else case when TEST.TEST_NATURE_LIVRAISON = 'E' then 'A ENLEVER' else null end end NATURE_LIVRAISON,
      case when TEST.TEST_MODE_VENTE = 'M' then 'MAGASIN' else case when TEST.TEST_MODE_VENTE = 'D' then 'DIRECT' else null end end MODE_VENTE,
      case when TEST.TEST_ETAT = 'I' then 'A RELANCER' else case when TEST.TEST_ETAT = 'R' then 'ABOUTI PARTIEL' else null end end ETAT_TESTIS,
      cast(datepart("year", TEST.TEST_DATE) * 10000 + datepart("month", TEST.TEST_DATE) * 100  + datepart("day", TEST.TEST_DATE) as varchar) DATE_TESTIS,
      cast(datepart("year", TEST.TEST_DATE_PREVUE_LIVR) * 10000 + datepart("month", TEST.TEST_DATE_PREVUE_LIVR) * 100  + datepart("day", TEST.TEST_DATE_PREVUE_LIVR) as varchar) DATE_VALIDITE,
      cast(datepart("year", TEST.TEST_DATE_SOUHAITEE) * 10000 + datepart("month", TEST.TEST_DATE_SOUHAITEE) * 100  + datepart("day", TEST.TEST_DATE_SOUHAITEE) as varchar) DATE_TRAVAUX,
      convert(varchar, TEST.TEST_MNT_TOT_HT_BON_PX_DEP) MONTANT_TOTAL_HT_PRIX_DEPART,
      convert(varchar, TEST.TEST_TOT_HT_FRAIS_TRANSP) TOTAL_HT_FRAIS_TRANSPORTS_PRODUITS
From      ENT_TESTIS as TEST, TECH_REGION as REGION, TECH_EXTRACT_CLOE as EXTRACT_CLOE, ENT_LIEN_ATC_CLI_ENCOURS as ACE
Where       (1=1)
And (REGION.TRE_ACTIF='O')
 And (TEST.TEST_TYPE IS NOT NULL AND TEST.TEST_TYPE = 'D' AND
TEST.TEST_ETAT in ('I','R') AND
(TEST.TEST_ECARTE_RELANCE IS NULL OR (TEST.TEST_ECARTE_RELANCE IS NOT NULL AND TEST.TEST_ECARTE_RELANCE != 'O')) AND
TEST.TEST_DATE_PREVUE_LIVR IS NOT NULL AND TEST.TEST_DATE_PREVUE_LIVR >= dateadd("month", -3, CURRENT_TIMESTAMP) AND
TEST.TEST_TECH_DATE_SUPP IS NULL AND TEST.TEST_TECH_DATE_SUPP_CLOE IS NULL)
 And (Coalesce(TEST.TEST_TECH_DATE_MAJ,TEST.TEST_TECH_DATE_CREA) > EXTRACT_CLOE.TEC_DATE_EXTRACT
OR
Coalesce(ACE.ACE_TECH_DATE_MAJ,ACE.ACE_TECH_DATE_CREA) > EXTRACT_CLOE.TEC_DATE_EXTRACT)
 And (EXTRACT_CLOE.TEC_GROUPE='TEST')
 And (ACE.ACE_ORIGINE_CLIENT = 'S')
 And (TEST.REG_CODE=REGION.TRE_CODE)
AND (TEST.REG_CODE=EXTRACT_CLOE.TEC_REG_CODE)
AND (TEST.SOC_CODE=ACE.SOC_CODE AND TEST.CLI_CODE=ACE.CLI_CODE)


Thanks
bibi
0
Comment
Question by:bibi92
  • 3
  • 2
  • 2
7 Comments
 
LVL 29

Expert Comment

by:QPR
ID: 33742639
have you considered the sql 2008 query optimizer
0
 
LVL 29

Expert Comment

by:QPR
ID: 33742654
0
 
LVL 50

Expert Comment

by:Lowfatspread
ID: 33743502
try this

what is the main current problem?

what indexes are available / being used?

how many rows are returned?
how many rows on each table?

are you certain you need the distinct?
 what order do you want the results returned in ?

 NB you do not have to specify
        x is not null and x <> 'B'
      since x <> 'B'   will ignore null values anyway




 
 
Select      Distinct      
      TEST.SOC_CODE CODE_SOCIETE,
      TEST.CLI_CODE CODE_CLIENT,
      TEST.TEST_ORIGINE_CLIENT CODE_ORIGINE_CLIENT,
      TEST.AGE_CODE CODE_AGENCE,
      TEST.TEST_CODE NUMERO_TESTIS,
      TEST.TEST_VERSION CODE_VERSION_TESTIS,
      case TEST.TEST_NATURE_LIVRAISON when 'L' then 'LIVRE' when 'E' then 'A ENLEVER' else null end NATURE_LIVRAISON,
      case TEST.TEST_MODE_VENTE when 'M' then 'MAGASIN' when 'D' then 'DIRECT' else null end MODE_VENTE,
      case TEST.TEST_ETAT when 'I' then 'A RELANCER' else when 'R' then 'ABOUTI PARTIEL' else null end ETAT_TESTIS,
      convert(char(8),TEST.TEST_DATE ,112) DATE_TESTIS,
      convert(char(8),TEST.TEST_DATE_PREVUE_LIVR,112) DATE_VALIDITE,
      convert(char(8),TEST.TEST_DATE_SOUHAITEE,112) DATE_TRAVAUX,
      convert(varchar, TEST.TEST_MNT_TOT_HT_BON_PX_DEP) MONTANT_TOTAL_HT_PRIX_DEPART,
      convert(varchar, TEST.TEST_TOT_HT_FRAIS_TRANSP) TOTAL_HT_FRAIS_TRANSPORTS_PRODUITS
From      ENT_TESTIS as TEST
inner join (select tre_code 
              from TECH_REGION 
             Where TRE_ACTIF='O'             
            )    as REGION
   ON TEST.REG_CODE=REGION.TRE_CODE
Inner Join (select tec_reg_code,tec_date_extract 
     from TECH_EXTRACT_CLOE
    Where TEC_GROUPE='TEST'
    Group by tec_reg_code,tec_date_extract    
   )    as EXTRACT_CLOE
 On TEST.REG_CODE=EXTRACT_CLOE.TEC_REG_CODE  
Inner Join (select Distinct soc_code,cli_code
                  ,coalesce(ace_tec_date_maj,ace_tech_date,crea) as Acedate 
     from ENT_LIEN_ATC_CLI_ENCOURS
    Where ace_origine_client='S' 
   )  as ACE
  On TEST.SOC_CODE=ACE.SOC_CODE
 AND TEST.CLI_CODE=ACE.CLI_CODE
   
Where 
  TEST.TEST_TYPE = 'D' 
       AND TEST.TEST_ETAT in ('I','R') 
       AND(TEST.TEST_ECARTE_RELANCE IS NULL               
             or TEST.TEST_ECARTE_RELANCE != 'O'
           )
       AND TEST.TEST_DATE_PREVUE_LIVR >= dateadd("month", -3, CURRENT_TIMESTAMP) 
       AND TEST.TEST_TECH_DATE_SUPP IS NULL 
       AND TEST.TEST_TECH_DATE_SUPP_CLOE IS NULL
     
 And (Coalesce(TEST.TEST_TECH_DATE_MAJ,TEST.TEST_TECH_DATE_CREA) > EXTRACT_CLOE.TEC_DATE_EXTRACT
       OR Ace.Acedate > EXTRACT_CLOE.TEC_DATE_EXTRACT
     )
 

Open in new window

0
Simplifying Server Workload Migrations

This use case outlines the migration challenges that organizations face and how the Acronis AnyData Engine supports physical-to-physical (P2P), physical-to-virtual (P2V), virtual to physical (V2P), and cross-virtual (V2V) migration scenarios to address these challenges.

 

Author Comment

by:bibi92
ID: 33744794
Hi,
200000 ROWS are returned
Indexes are created on PK.
Tables are nearly 1000000rows.
The use of distinct is necessary for the customer.
The order is not indispensable.

Thanks
bibi
0
 
LVL 50

Accepted Solution

by:
Lowfatspread earned 500 total points
ID: 33751789
1) are all 200K rows required?
2) what is actually making the rows not distinct?
          1:m relationships... , removal of the time component of the dates, ....?

3) Do all tables contain 1 million rows?
      if not what are the actual table sizes

4) what are the tables primary keys?
    are the PK indexes clustered?
    any reason these are the only indexes?
    how are the tables used... (what sort of system is this .. OLTP,MIS,OLAP, mixed, ...)?

 
 
Select      Distinct      
      TEST.SOC_CODE CODE_SOCIETE,
      TEST.CLI_CODE CODE_CLIENT,
      TEST.TEST_ORIGINE_CLIENT CODE_ORIGINE_CLIENT,
      TEST.AGE_CODE CODE_AGENCE,
      TEST.TEST_CODE NUMERO_TESTIS,
      TEST.TEST_VERSION CODE_VERSION_TESTIS,
      case TEST.TEST_NATURE_LIVRAISON when 'L' then 'LIVRE' when 'E' then 'A ENLEVER' else null end NATURE_LIVRAISON,
      case TEST.TEST_MODE_VENTE when 'M' then 'MAGASIN' when 'D' then 'DIRECT' else null end MODE_VENTE,
      case TEST.TEST_ETAT when 'I' then 'A RELANCER' when 'R' then 'ABOUTI PARTIEL' else null end ETAT_TESTIS,
      convert(char(8),TEST.TEST_DATE ,112) DATE_TESTIS,
      convert(char(8),TEST.TEST_DATE_PREVUE_LIVR,112) DATE_VALIDITE,
      convert(char(8),TEST.TEST_DATE_SOUHAITEE,112) DATE_TRAVAUX,
      convert(varchar, TEST.TEST_MNT_TOT_HT_BON_PX_DEP) MONTANT_TOTAL_HT_PRIX_DEPART,
      convert(varchar, TEST.TEST_TOT_HT_FRAIS_TRANSP) TOTAL_HT_FRAIS_TRANSPORTS_PRODUITS
From      ENT_TESTIS as TEST
Inner Join (select a.tec_reg_code,a.tec_date_extract 
              from TECH_EXTRACT_CLOE as a
             Inner Join TECH_REGION as b  
                on a.tec_reg_code=b.tre_code
             Where a.TEC_GROUPE='TEST'
               and b.TRE_ACTIF='O'
             Group by a.tec_reg_code,a.tec_date_extract    
           )    as EXTRACT_CLOE
 On TEST.REG_CODE=EXTRACT_CLOE.TEC_REG_CODE  
Inner Join (select Distinct soc_code,cli_code
                  ,coalesce(ace_tec_date_maj,ace_tech_date,crea) as Acedate 
     from ENT_LIEN_ATC_CLI_ENCOURS
    Where ace_origine_client='S' 
   )  as ACE
  On TEST.SOC_CODE=ACE.SOC_CODE
 AND TEST.CLI_CODE=ACE.CLI_CODE
   
Where 
  TEST.TEST_TYPE = 'D' 
       AND TEST.TEST_ETAT in ('I','R') 
       AND(TEST.TEST_ECARTE_RELANCE IS NULL               
             or TEST.TEST_ECARTE_RELANCE != 'O'
           )
       AND TEST.TEST_DATE_PREVUE_LIVR >= dateadd("month", -3, CURRENT_TIMESTAMP) 
       AND TEST.TEST_TECH_DATE_SUPP IS NULL 
       AND TEST.TEST_TECH_DATE_SUPP_CLOE IS NULL
     
 And (Coalesce(TEST.TEST_TECH_DATE_MAJ,TEST.TEST_TECH_DATE_CREA) > EXTRACT_CLOE.TEC_DATE_EXTRACT
       OR Ace.Acedate > EXTRACT_CLOE.TEC_DATE_EXTRACT
     )
 

Open in new window

0
 

Author Comment

by:bibi92
ID: 33765886
Hello,

Thanks

1) are all 200K rows required? ---> Yes
2) what is actually making the rows not distinct? ---> dates
          1:m relationships... , removal of the time component of the dates, ....?

3) Do all tables contain 1 million rows? ---> Yes
      if not what are the actual table sizes

4) what are the tables primary keys?
    are the PK indexes clustered? ---> Yes
    any reason these are the only indexes? I don't know. The customer can't answer to this question
    how are the tables used... ---> OLTP

Regards
bibi
0
 

Author Closing Comment

by:bibi92
ID: 33952962
Thanks bibi
0

Featured Post

Free Tool: IP Lookup

Get more info about an IP address or domain name, such as organization, abuse contacts and geolocation.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

There have been several questions about Large Transaction Log Files in SQL Server 2008, and how to get rid of them when disk space has become critical. This article will explain how to disable full recovery and implement simple recovery that carries…
Occasionally there is a need to clean table columns, especially if you have inherited legacy data. There are obviously many ways to accomplish that, including elaborate UPDATE queries with anywhere from one to numerous REPLACE functions (even within…
Although Jacob Bernoulli (1654-1705) has been credited as the creator of "Binomial Distribution Table", Gottfried Leibniz (1646-1716) did his dissertation on the subject in 1666; Leibniz you may recall is the co-inventor of "Calculus" and beat Isaac…
Attackers love to prey on accounts that have privileges. Reducing privileged accounts and protecting privileged accounts therefore is paramount. Users, groups, and service accounts need to be protected to help protect the entire Active Directory …

735 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question