0EQUIPMENT_ATTR taking too long to finish

[Former Member]

Dear All,

I have a full load IP for datadource 0EQUIPMENT_ATTR, which is already divided in to parts based on date. One of the part, which extracts data from 2013 to 2999 is taking too much time to complete ~18 hours. This has recently occurred after patch upgrade to SP11

Indexes are already created on the required table.

I have checked the corresponding job in ECC system and based on my analysis(may be wrong) it seems that system is taking time to call the Customer enhancement. See Below screenshot.

Are there any notes or any suggestions that can prove helpful in this case?

Regards,

AR

[Former Member · Accepted solution]

Hello All,

The issue is resolved. There was a small piece of code that was taking time.It was commented in Dev and QA but somehow after upgrade it was active in PRD

Reg,

Amit

[Former Member]

Hi Amith,

We also got the same issue after upgrade on 0EQUIPMENT_ATTR; we have raised an
Oss message to SAP. The solution as below

The basic cause of the old behavior were Oracle-DB specific settings
in the DB statistics (of indexes).
A colleague of mine adjusted these from value 4883 to 5189 for column
ERDAT and index EQUI-ZB1. This result in Oracle using a different Join
strategy with presence on table EQUI rather than EQUZ

Please check the incident 20734 year 2014.

Regards,

polu.

[Former Member]

Hello Pola,

The INC does not exists anymore.

Can you tell me more about the resolution part?

Ragards,

Amit

[Former Member]

Hi Amit,

I would say this must be due to amount of data in the EQUI table. Can you check how many records do you have in view for your date range.

The region you have highlighted in job log is the time datasource extracts data from view BIW_EQUI and then after selecting data it then goes to check BTE. So from 02:06:52 it starts selection from table and completes the selection in about 1 minute when it calls BTE.

Also it is processing around 1000 records in one package. So if you have huge amount of data then it will take longer.

Was it not taking this long before upgrade for same time period? I am assuming that you are applying pseudo delta for this on both dates AEDAT and ERDAT. I fnot then I suggest you write a routine in infopackage selection and extract only 1 day of data by performing pseudo delta (FULL selection based).

Can you attach the complete job log if above does not help?

Thanks

Amit

[Former Member]

Hello Amit,

Yes there are million(8 mil) of records this IP is pulling. But before upgrade this used to complete in 1 hour max. Now its taking 18 hours everyday.

I will try and use the pseudo delta option you have mentioned. But we are loading FULL data as per business requirement. So cant say anything for sure.

The full job log is same as i have attched in the main question above.

IS it possible to apply index on a view?

Regards,

AR