Performance problem with EDI40 while unicode conversion

[Former Member]

Hi All,

We are doing Unicode migration of our BI 7.0 EHP1 SP 6 system. However we have completed must of the export in 24 Hr. ; but still one R3load job is running since from last 2 days i.e. more than 48 Hr. It is extracting the data from cluster table i.e. EDI40.

As concern to this issue I have checks the SAP Note 872124, 991831, but we are using right R3load version i.e. R7.01/V1.4 Compiled Feb 24 2009 23:50:52.

So please help me to reduce this conversion time, as it is crossing our downtime limit.

Apart from this I have check the SAP Note 706478, as per note we need to do the archiving of old entries from EDI40 table. But we required following information before we start archiving / deletion of old entry from EDI40 table in source system.

1. Do we require these old entries in target system, as concern to normal system option?

2. Is ok if we truncate this table in source system before starting Unicode conversion?

Following is our system and EDI40 table information.

SAP System: BI 7.0 with EHP1 SP6

Database: MS SQL 2005

OS: Windows 2003

Database Size : 2.4 TB

Hardware information :

CPU : 16 CUP ( Intel 1598 Mhz)

RAM : 65 GB

EDI40 information:

Data Space 20,206.453 MB

Index Space 196.289 MB

Row count 10567724

Thanks,

Hatshal

[nils_buerckel]

Hi Harshal,

did you already consider to use table splitting for EDI40 ?

Best regards,

Nils Buerckel

SAP AG

[Former Member]

Hi Nils,

Thanks a lot for your reply.

Yes, I tried to do the table splitting for EDI40 table but it was taking some much time. So I have cancelled that and started the unicode conversion.

But can you please provide some information on following.

1.As per SAP Note 706478, if we do the archiving / deletion of old entries from EDI40 table. Then do we need these old entries in target system, as concern to normal system option?

2. OR Is ok if we truncate this table in source system before starting Unicode conversion?

Thanks,

Harshal

[nils_buerckel]

Hi Harshal,

please check also SAP note 40088.

1. If you archive the objects, they should be available (as display) in the Unicode system as well - otherwise there is no difference to deletion.

2. EDI40 contains IDocs. As far as I know it, completed IDocs can be important for auditor purposes. Therefore I do not think that it makes sense to truncate this table !

Best regards,

Nils Buerckel

SAP AG