La vue SQL génère les lignes en double

Question

Je crée une nouvelle requête à partir d'une ancienne requête et j'obtiens des lignes en double dans mon résultat. Je n'ai absolument aucune idée de pourquoi cela se passe?

Voici les anciens codes SQL qui ne produisent pas de doublons:

XtraViwStock

SELECT  TOP 100 PERCENT dbo.qexp_detail.in_date AS inDate, RTRIM(dbo.qexp_detail.stock_locn) AS stockLocn, { fn LCASE(RTRIM(dbo.qexp_detail.in_locn)) 
         } AS inLocn, dbo.qin_doc.in_doc_no AS inDoc, RTRIM(dbo.qexp_detail.client_plt_id) AS pltId, dbo.xtraviwItem.cm, dbo.xtraviwItem.vari, 
         dbo.xtraviwItem.pack, dbo.qexp_detail.prod_grp AS pGrp, dbo.qexp_detail.prod_char AS pChar, dbo.xtraviwItem.grd, dbo.xtraviwItem.sizeCount, 
         dbo.xtraviwItem.brand, dbo.xtratblNotifPltId.chrNotifPltId AS notifPltId, dbo.xtraviwItem.ic, dbo.qexp_detail.targ_mkt AS tm, dbo.qexp_detail.farm AS puc, 
         dbo.qexp_detail.ctn_qty AS ctns, CAST(dbo.qexp_detail.ctn_qty AS float)/CAST(dbo.xtraviwCtnsPerPlt_st.cpp AS float) AS plts, 
         dbo.xtraviwCtnsPerPlt_st.cpp, dbo.qexp_detail.plt_base AS base, dbo.xtraviwMessage.msgText AS inDocMsg, dbo.xtratblSpec.intSpecId AS spec1Id, 
         dbo.xtratblSpec.strInstruction AS spec1Instruct, dbo.xtratblSpec.dtmSpecDate AS spec1SpecDate, 
         dbo.xtratblSpecPallet.dtmLoadDate AS spec1LoadDate, dbo.xtratblSpecPallet.strLoadLocn AS spec1LoadLocn, 
         dbo.xtratblSpecPallet.strReceiver AS spec1Rec, dbo.xtratblSpecPallet.strPOD AS spec1Pod, dbo.xtratblSpec2.intSpecId AS spec2Id, 
         dbo.xtratblSpec2.strInstruction AS spec2Instruct, dbo.xtratblSpec2.dtmSpecDate AS spec2SpecDate, 
         dbo.xtratblSpecPallet2.dtmLoadDate AS spec2LoadDate, dbo.xtratblSpecPallet2.strLoadLocn AS spec2LoadLocn, 
         dbo.xtratblSpecPallet2.strReceiver AS spec2Rec, dbo.xtratblSpecPallet2.strPOD AS spec2Pod, dbo.xtraviwItem.org, dbo.qexp_detail.tran_user, 
         dbo.qexp_detail.tran_datetime, 
         dbo.xtratblSupplier.strProperName + ' ' + dbo.xtratblSupplier.strProperSurname + ', ' + dbo.xtratblSupplier.strProperFarms + ', ' + dbo.qexp_detail.supplier 
         AS producer, { fn LCASE(RTRIM(dbo.qin_doc.farm)) } AS packLocn, dbo.xtraviwItem.varietyGroup, dbo.xtraviwItem.brandDesc, dbo.xtraviwItem.kg, 
         dbo.qexp_detail.client_plt_type AS pltHt, dbo.xtraviwItem.hcHt, dbo.xtraviwItem.stdHt, dbo.qexp_detail.build_flag AS buildFlag 
FROM   dbo.xtraviwCurrentSeason INNER JOIN 
         dbo.qexp_detail ON dbo.xtraviwCurrentSeason.season = dbo.qexp_detail.season LEFT OUTER JOIN 
         dbo.qsupplier LEFT OUTER JOIN 
         dbo.xtratblSupplier ON dbo.qsupplier.supplier = dbo.xtratblSupplier.chrSupplier ON dbo.qexp_detail.supplier = dbo.qsupplier.supplier LEFT OUTER JOIN 
         dbo.xtratblSpec2 RIGHT OUTER JOIN 
         dbo.xtratblSpecPallet2 ON dbo.xtratblSpec2.intSpecId = dbo.xtratblSpecPallet2.intSpecId ON 
         dbo.qexp_detail.client_plt_id = dbo.xtratblSpecPallet2.chrPltId AND dbo.qexp_detail.season = dbo.xtratblSpecPallet2.intSeason LEFT OUTER JOIN 
         dbo.xtratblSpec RIGHT OUTER JOIN 
         dbo.xtratblSpecPallet ON dbo.xtratblSpec.intSpecId = dbo.xtratblSpecPallet.intSpecId ON 
         dbo.qexp_detail.client_plt_id = dbo.xtratblSpecPallet.chrPltId AND dbo.qexp_detail.season = dbo.xtratblSpecPallet.intSeason LEFT OUTER JOIN 
         dbo.qfarm ON dbo.qexp_detail.farm = dbo.qfarm.farm LEFT OUTER JOIN 
         dbo.xtraviwMessage RIGHT OUTER JOIN 
         dbo.qin_doc ON dbo.xtraviwMessage.msgNr = dbo.qin_doc.mesg_no ON dbo.qexp_detail.in_load_id = dbo.qin_doc.in_load_id LEFT OUTER JOIN 
         dbo.xtraviwItem ON dbo.qexp_detail.item_code = dbo.xtraviwItem.item LEFT OUTER JOIN 
         dbo.xtratblNotifPltId ON dbo.qexp_detail.season = dbo.xtratblNotifPltId.intSeason AND 
         dbo.qexp_detail.client_plt_id = dbo.xtratblNotifPltId.chrDiparPltId LEFT OUTER JOIN 
         dbo.xtraviwCtnsPerPlt_st ON dbo.qexp_detail.season = dbo.xtraviwCtnsPerPlt_st.cppSeason AND 
         dbo.qexp_detail.season_key = dbo.xtraviwCtnsPerPlt_st.cppCm AND dbo.qexp_detail.client_plt_id = dbo.xtraviwCtnsPerPlt_st.cppPltId 
WHERE  (dbo.qexp_detail.plt_status = 'S') 
ORDER BY dbo.xtraviwItem.cm, dbo.qexp_detail.in_date, RTRIM(dbo.qexp_detail.stock_locn), RTRIM(dbo.qexp_detail.client_plt_id), dbo.qexp_detail.farm 

Et voici la nouvelle version qui crée des doublons:

AllStock2010

SELECT  TOP 100 PERCENT dbo.qexp_detail.in_date AS InDate, dbo.qexp_detail.stock_locn AS StockLocn, dbo.qexp_detail.in_locn AS InLocn, 
         dbo.qin_doc.in_doc_no AS InDoc, dbo.qxfer_detail.instruction_no AS TransferInstructNo, dbo.qexp_detail.instruction_no AS LoadInstructNo, 
         dbo.qexp_detail.client_plt_id AS PltID, dbo.xtraviwItem.cm AS CM, dbo.xtraviwItem.vari AS Variety, dbo.xtraviwItem.pack AS Pack, 
         dbo.qexp_detail.prod_grp AS ProdGrp, dbo.qexp_detail.prod_char AS ProdChar, dbo.xtraviwItem.grd AS Grade, 
         dbo.xtraviwItem.sizeCount AS SizeCount, dbo.xtraviwItem.brand AS Brand, dbo.xtraviwItem.ic AS IC, dbo.qexp_detail.targ_mkt AS TM, 
         dbo.qexp_detail.farm AS PUC, dbo.qexp_detail.ctn_qty AS Cartons, CAST(dbo.qexp_detail.ctn_qty AS float)/CAST(dbo.xtraviwCtnsPerPlt_st.cpp AS float) 
         AS Plts, dbo.xtraviwCtnsPerPlt_st.cpp AS CpP, dbo.qexp_detail.plt_base AS PltBase, dbo.xtraviwMessage.msgText AS InDocMsg, 
         dbo.xtraviwItem.org AS Org, dbo.qexp_detail.tran_user, dbo.qexp_detail.tran_datetime, dbo.qexp_detail.supplier AS Producer, 
         dbo.qin_doc.farm AS PackLocn, dbo.qin_detail.pack_date AS PackDate, dbo.xtraviwItem.varietyGroup AS VarietyGroup, 
         dbo.xtraviwItem.brandDesc AS BrandDesc, dbo.xtraviwItem.kg AS Kg, dbo.qexp_detail.client_plt_type AS PltHt, 
         dbo.qexp_detail.build_flag AS BuildFlag, dbo.xtraviwItem.hcHt AS HcHt, dbo.xtraviwItem.stdHt AS StdHt, dbo.qexp_detail.plt_status AS PltStatus 
FROM   dbo.xtraviwItem RIGHT OUTER JOIN 
         mynhardt.CurrentSeason INNER JOIN 
         dbo.qexp_detail ON mynhardt.CurrentSeason.Season = dbo.qexp_detail.season ON 
         dbo.xtraviwItem.item = dbo.qexp_detail.item_code LEFT OUTER JOIN 
         dbo.qxfer_detail ON dbo.qexp_detail.client_plt_id = dbo.qxfer_detail.client_plt_id LEFT OUTER JOIN 
         dbo.qin_detail ON dbo.qexp_detail.client_plt_id = dbo.qin_detail.client_plt_id LEFT OUTER JOIN 
         dbo.xtraviwCtnsPerPlt_st ON dbo.qexp_detail.season = dbo.xtraviwCtnsPerPlt_st.cppSeason AND 
         dbo.qexp_detail.season_key = dbo.xtraviwCtnsPerPlt_st.cppCm AND 
         dbo.qexp_detail.client_plt_id = dbo.xtraviwCtnsPerPlt_st.cppPltId LEFT OUTER JOIN 
         dbo.qin_doc ON dbo.qexp_detail.in_load_id = dbo.qin_doc.in_load_id LEFT OUTER JOIN 
         dbo.xtraviwMessage ON dbo.qin_doc.mesg_no = dbo.xtraviwMessage.msgNr 
WHERE  (dbo.qexp_detail.plt_status = 'S') 
ORDER BY dbo.xtraviwItem.cm, dbo.qexp_detail.in_date, dbo.qexp_detail.stock_locn, dbo.qexp_detail.client_plt_id, dbo.qexp_detail.farm 

J'espère que quelqu'un peut me diriger dans la bonne direction!

Answer 1

Je ne pense pas qu'il soit possible pour quiconque de résoudre cela pour vous.

Peut-être que cela peut être un début:
Remplacer les colonnes de votre nouvelle SELECT par un COUNT(*) et commencer commentant JOIN s jusqu'à ce que les changements de comptage retournés. Vérifiez si le JOIN que vous venez de commenter a causé que les lignes en double vous ne vous attendiez pas et assurez-vous que les conditions pour ce JOIN sont corrects.

Answer 2

GAUCHE/INNER JOIN sont mélangés avec RIGHT JOIN - à moins que ces tables ont une à une relation, vous aurez les doublons

Votre « plan de bataille » serait la suivante:

1. Vous devez dessiner un schéma pour toutes les tables impliquées et leur relation (1-à-plusieurs, etc.) et déterminer quelle (s) table (s) est votre source principale et quelles tables sont des tables de 'recherche'.

2. Supprimer toutes les tables de consultation (première et leurs champs inclus dans ensemble de données résultant), puis vérifier les doublons

3. Commencer à ajouter des tables de consultation, un par un chacun pour vérifier les doublons.

De plus, "TOP 100 PERCENT" est vraiment tout ensemble de données résultant.