Calculation View Activation time

[Former Member]

Hello,

When I activate a calculation view that is being used in many places it takes hours for it to activate and my workstation is running out of CPU and memory. If I copy the same calculation view aside and activate it the activation takes seconds (I guess its because the copy is not being used anywhere…)

Can someone please explain what does the system does during the activation and why it uses my workstation resources? Is it possible to make the activation faster?

Thanks,

Amir

[former_member184768 · Accepted solution]

If you are trying to activate a view which is used in another views, then all those dependent views might get activated. That you can see from the activation log.

If any of the dependent views have any errors / inconsistency then the system will try to validate those views which could take a longer time.

Suggest you to check the "where used" list for the given calc view and check the status of the dependent views. Delete the views which are no longer required. That will bring down the activation time.

Regards,

Ravi

[Former Member · Accepted solution]

Thanks Ravi,

I can see all the views that it is activating in the activation log - this must be the reason for slow activation time.  If i remember correctly in older releases the activation was much faster and I think it wasn’t checking all the views – is there a way to activate ionly the views without the dependencies?

Is there a place of log that other developers can see my job log as well (status of activation)?

Why does my workstation CPU and memory so high? If working from WEB IDE will make it faster?

Thanks,

Amir

[former_member184768 · Accepted solution]

Based on my knowledge:

1) Activation logs are available in your local workstation HANA Studio home. Search for the directory com.sap.ndb.studio.bi.logs, it should have the activation logs. You can find the information from HANA Studio, go to windows -> Preferences -> SAP HANA -> Modeler -> Logs -> Job logs. It will show the job log directory.

2) Secondly, when you activate a view, not only the dependent views, but it may go in recursive mode to redeploy the subsequent dependent views too. You can check that from the job log itself.

So far I couldn't figure out how to prevent this behavior. Neither do I wish to prevent this behavior as I want all the impacted (level 1, 2, 3 ... ) views get redeployed to avoid any inconsistency due to the change made to the base view.

Regards,

Ravi

[Former Member]

Hi Amir,

You can try the following. In the calc view properties, the default value for 'Data Category' will be 'CUBE'. Try to make it as Blank.This will prevent generation of column views for each column of the calc view.

When this will help ->

This can for surely be made as blank for the calculation views that are not used for reporting or if it is used for reporting by purely SQL based tools.So it can be retained only in the top level calc views that are used to build queries and for all underlying calc views the datacategory can be blank.

We had taken such a approach and were able to reduce the activation runtime drastically in dev as well as during further transports.

When this will not help->

The setting needs to be CUBE if this calc view is used for eg. in ODP datasource or in virtual providers. Maybe there could be other cases where this setting should not be blank if using business objects frontend etc that you may need to watch out for.

Regards,

Mandeep