Reverse Engineering from Oracle without bringing in Partitions

[former_member1194361]

I am using PD 16.5 SP02

When you database compare an existing Physical data model or Reverse Engineer one, PD gives no option to uncheck bringing in Actual physical partitions and bring just the default MAX/MAXVALUE partition.

I have some tables which are daily partitioned with partitions created for 2 years, so about 730 partitions. Even on a Core i5 machine with 8 GB RAM. PD chokes trying to first reverse engineer or database compare. Then it chokes while you try to right click a table and select Physical Options -> Partitions. It takes really long time to do any of these functions. And then if you try to compare such a partition full rev engineered data model to a production data model sitting in repository (with an intention of syncing the model to what is in the database), it then takes really long time to show the compare results screen.

Anyone else faced similar situation when working with this version of PD and Oracle 11/12

[marc_ledier]

Hi,

You can untick the checkbox "Physical Options" on the Table page of the reverse engineering.

This should accelerate the reverse.

Regards,

Marc

[former_member200945]

This is a known issue.

Do you have to generate these partitions?

[former_member1194361]

No, I do not need to generate the "real" partitions back. I just need to save in the data model the fact that this table is partitioned. In the normal workflow where the table was created from a DDL generated out of data model, we create just the MAX/MAXVALUE partition and then the DBA creates the real partitions, depending on whether it is monthly partitioned or daily partitioned and the retention period.

But PD does not give either of the below options when reverse engineering from an Oracle database

1. Get only the default partition (i.e MAX/MAXVALUE)

2. Do not get any partition info

I feel both of these should be options in a future release. Today if you reverse engineered a daily partitioned table with 2 years retention, it pulls in 730 or so partitions and then takes really long to open the physical options -> partitions window and also takes really wrong if this reverse engineered data model was complete compared with a production data model in repository.

[former_member200945]

See  https://scn.sap.com/thread/3892249

[former_member1194361]

That link points right back to this thread, were you trying to post something else ?

[former_member200945]

Performance issue with elaborate partition defi... | SCN

[former_member185199]

What is funny (in a sarcastic manner) (and also shows the QUALITY of the Software!)

is that the article you mentioned is 2 years old