X hits on this document

97 views

0 shares

0 downloads

0 comments

31 / 48

Microsoft’s SAP ERP (R/3) System (cont)

Multiple Problems with multiple solutions in this scenario:

When query is cached, dependent on range, users might experience undesirable performance:

Small ranges might be executed over clustered index scans

Wide ranges might be executed over non-clustered index scan in other cases

Solutions:

Change attribute of clustered index to non-clustered and make former non-clustered index on columns Client and Date to clustered index

Only works when columns in Primary Key (like Ordernumber) are not including semantics which can be used to range scan on. Some ISVs meanwhile use GUIDs as Ordernumbers. GUIDs never should be included in clustered indexes

Add recompile hint to query

DBA-416M• ISV configurations & implementations – Lessons Learned

Document info
Document views97
Page views97
Page last viewedTue Dec 06 10:32:52 UTC 2016
Pages48
Paragraphs862
Words4327

Comments