SAP CRM Fiori 应用 my Opportunities 的后台性能分析

The performance bottleneck of this url /sap/opu/odata/NBA/CRM_OPPORTUNITY_SRV/Opportunities()?KaTeX parse error: Expected 'EOF', got '&' at position 99: …ode eq 'E0003')&̲orderby=Score desc&KaTeX parse error: Expected 'EOF', got '&' at position 7: top=50&̲select=Id,CIFId,PriorityCode,SalesStageText,StartDate,ClosingDate,ProcessTypeText,OpportunityTypeText,EmployeeResponsibleName,ProspectShortName,OpportunityType,Description,Guid,UserStatusCode,ProcessType,Score,RelationshipView,Changed_At,Created_At

is in one order FM CRM_BSP_OIC_1O_SEARCH_FROM_RF itself. I have captured all the importing parameters for it and done an SAT trace in NR1/090, finally the bottleneck ( more than 3 seconds ) is found in the OPEN SQL below:

In the OPEN SQL there are inner joins on five large table:

CRMD_ORDER_INDEX  INNER JOIN CRM_JEST AS CRM_JEST_H ON CRM_JEST_H~OBJNR = CRMD_ORDER_INDEX~HEADER INNER JOIN CRMD_ORDERADM_H ON CRMD_ORDERADM_H~GUID = CRMD_ORDER_INDEX~HEADER  INNER JOIN CRMD_LINK ON CRMD_LINK~GUID_HI = CRMD_ORDER_INDEX~HEADER INNER JOIN CRMD_PARTNER ON CRMD_PARTNER~GUID = CRMD_LINK~GUID_SET

The table entries for each table are listed below:

CRMD_ORDER_INDEX: 25,694,983
CRM_JEST: 31,673,473
CRMD_ORDERADM_H:7,627,583
CRMD_LINK:45,508,977
CRMD_PARTNER: 27,277,824

Such inner join had poor performance and unfortunately, the OPEN SQL is populated dynamically via one order framework and could not be controlled by the application.

更多Jerry的原创文章，尽在：“汪子熙”: