how to deceide that v should load to ods or cube?

[Former Member]

hello all

having a doubt like how to deceide whether v should load data from r/3 to ods or cube,on what basis should v deceide?

Many Thanks

balaji

[Former Member · Accepted solution]

Hi Balaji,

Here are few points:

1. For detailed level of data use ODS.

2. If the extractor is not supporting delta go for ods from which u can load in to other data targets.

3. Unlike with InfoCubes, you can also choose to overwrite fields. This is especially important when it comes to document-type structures; if you change documents in the OLTP; the changes not only cover numeric fields such as order quantity, but also non-numeric fields such as goods receiver, status and delivery date. So that these changes can also be copied in the BW ODS Objects, the relevant fields in the ODS Objects must also be overwritten and set to the current value.

Regards,

Amit.

Message was edited by: Amit

[Former Member · Accepted solution]

Amit

Thanks for your valuable info given?

can u give me example on this what u said like "If the extractor is not supporting delta go for ods from which u can load in to other data targets".

Many Thanks

balaji

[Former Member]

Hi Balaji,

If you want to have unique records for ODS. You have to Check the setting of the ODS . Unique Data Records.

The scenario can be an R/3 Field ( which is not a Key Field ) can be made as a Key field in your ODS but you wish to have it as unique, in this case it helps.

In flat file loading also this helps greatly.

Thanks,

Srinivas.

[Former Member]

Hi Balaji,

Some of the features for which you can use an ODS are

• For delta enabling the data, if the R/3 datasource does not give you delta updates and you have to extract full uploads.

• If you need to have detail or line item data available for indepth analysis even though you might be reporting on aggregated data from an Infocube.

If you are looking for any of the above features then you would consider getting your data to/through an ODS, else it would be advisable to load data to the cube for reporting purposes as the cube is designed on the principles of Data warehousing and hence suited for reporting. Whereas and ODS is just a flat file structure and hence not suitable for reporting of large data sets.

Hope it helps.

Anurag

[Former Member]

Besides the good elements already provided, from a genuine technical viewpoint the difference is mainly this one: do you need to use the key concept or don't ?

In other words, if you need to have on only one row a set of info coming from several flows (sales, delivery and billing), well, you definitely need an ODS !

Hope it helps!

Bye,

Roberto

[Former Member]

Roberto

Thanks for your reply,in your statement u said like "if you need to have on only one row a set of info coming from several flows (sales, delivery and billing), well, you definitely need an ODS",is there any specific reason behind?I mean how it helps out by loading all this three(sales, delivery and billing)datsource to ods?

Pl let me know

Many thanks

balaji

[Former Member]

If you want to see for a specific sales order document: order qty, delivered qty, billed qty, set specific status by comparing several fields coming from the different flows and so on...you need this kind of scenario!

(it's really a very common requirement, especially in Logistic, where different flows are related to the same business process...)

Bye,

Roberto

[Former Member]

hi Anurag

thanks for your detail info given,i need to have some clarification on your statements like "if the R/3 datasource does not give you delta updates and you have to extract full uploads",can u give me example on this part?

Many THanks

balaji

[Former Member]

Roberto

this type of info i needed,its really satisfying me a lot.

Thanks a lot

Many Thanks

balaji

[Former Member]

Hi Balaji,

For an example of R/3 datasource not providing delta updates, we had a scenario wherein we were extracting data from Datasource 0CO_OM_CCA_1, but it did not provide delta capability.

So we introduced an ODS before the final data was stored in the Infocube. As a result only changed records were passed on to the Infocube from the ODS.

Hope it helps.

Regards

Anurag

[Former Member]

Hi Balaji,

You can decide whether u have Unique Records By ( Enabling the settings in ODS) from R/3.

BAsed on whether You need to Pull data further from one or more First Level ODs To second Level ODS or Cube.

Eg: R/3 to ODS1

R/3 to ODS2

Then ODS1 + ODS2 to ODS3

ODS1 + ODS2 to Cube.

Thanks,

Srinivas.

[Former Member]

hi srinivas

thanks a lot four info given

can u give me example regarding in your statements "You can decide whether u have Unique Records By ( Enabling the settings in ODS) from R/3".

many thanks

balaji