DTP loading to InfoCube taking SO MUCH TIME

[fthomazetti_costa]

Hello Experts,

I know this is a recurrent question here, but I think that my problem has more specific reasons.

All loads to InfoCubes, not matter the size, no matter how many records, no matter if it from a DSO or a PSA or an InfoSet... ALWAYS take a lot of time, but I'm meaning A LOT OF TIME!

I saw in some forums, people angry because their DTP load to an InfoCube is taking 30 minutes to load around 15 millions of records. Man, I can't do this kind of load because it will take a entire life!

As I've said before, I've problem with ALL my InfoCube loads. Loads from InfoPackages or DTP load to an DSO is acceptable. Not that fast, but not that slow. Is fine for me.

If I've around 100 records, it'll take around 1 hour. If I've around 1 million, it'll take more than 18 hours.

I've already check my backgrounds, the package size, the parallel jobs and the star schema of the InfoCube, but no matter what I do, my InfoCube loads ALWAYS take a lot of time. Delta loads, full loads... very, very, very slow.

Maybe there's some parameter or configuration in SAP that I've to do, but I don't have any idea about it.

There is another thing that can I do?

Thanks,

[former_member242986]

HI Flavio,

Are there already data in the cube?

Did you try to delete the Indexes on that cube? (Independent if you created them, try to delete them before loading)

Regards

Jörg

[Former Member]

Hi Flávio,

You may trace the execution and then you can analyze the result:

1959493 - How to get ST12 trace for DTP request

or

2436955 - Step by step instructions on how to use ST12 trace for analysis

Best regards, Andras

[FCI]

Hi Flavio,

If you take the 1 hour load for 100 records, can't you see what is doing the system during this hour ? Any long time reading a specific table (should be visible with a SM50) ?

How old is your system ?

What you have described makes me think to a system table that is becoming too big and need to be emptied.

Regards,

Frederic

[fthomazetti_costa]

Hi Frederic,
The system looks good. I work in a huge international company, so system and hardware isn't the problem. I'm doing a load of around 400.000 records right now. This is the first load of this InfoCube, but the second tentative. The first one toke more than 9 hours and ended with errors due some special character in the field "text". I've included this characters on RSKC and I'm trying to load again and yes, first I've cleared the InfoCube, deleted the Index and the dimensions table. There's no other activities being execute by the system right now. Take a look at the SM50:

Runtime is taking so long as I've said before, but the progress is running as a normal condition.

And it is important to say, this is one guy that I'm executing right now... But this behavior and this "long running" always occurs for all the InfoCubes.

Regards,

Flávio