We are able to find out why .whr files generated by table splitter not picked up by R3load. The reason is first of all STR files not generated for these splitted tables by R3ldctl, the issue was resolved after giving the input file for all of these files under Predefined tables to be Extracted option in Package splitting.
I have few more questions. How can I control option 2 ?
Question 1 -
option 1 - Largest tables in separate packages ( partition tables are not considered for BI system ) - Picks up top tables based on size and input values excluding partition tables.
option 2 - Predefined tables to be Extracted with input file - Works fine as I am controlling it and can give partition tables here.
The question is does Option 2 still works if I don't select option 1 in package splitting tool ? I mean check mark box.
Question 2 -
We tried unsorted for one of the big partition table without split that took around 11 hours (/BIC/BXXXX - 150 Gig ) Vs sorted with table spitting. But to our surprise both took more or less equal amount of time. Our assumption was unsorted will cut down total amount of time by half i.e around 5 hours. Then what is the use of unsorted option then except the no usage of temp tablespace in oracle.
Question 3 -
While doing unsorted export, we have maintained ddlFile and orderBy text file with all packages except the two which are supposed to be exported in sorted order. But R3load still considered those two packages which I have not maintained in orderBy text file.
orderBy= OrderBy.txt
ddlFile = /export/exp_640/ABAP/DB/DDLORA_LRG.TPL
ddlMap =
I specified top 10 tables at the top then packages next but during export those tables are considered at the last.
Is it the correct way or am I missing something here ? Do I need to specify ddlMap for this ?
Appreciate your help and valuable information in this regard.