Semantic grouping in Transformations in SAP BW 7.3

[Former Member]

Hi

I have read that semantic grouping is possible in transformations in BW 7.3 just like in DTP's. As per my understanding, semantic grouping in DTP is to group the data records as per the key, upload them into data packets into the target. What is the purpose of semantic grouping in transformation and how can it be done?

Thanks

[Former Member · Accepted solution]

HI

Have a look at the below link

http://help.sap.com/saphelp_nw73/helpdata/en/4a/1df3235c171b40e10000000a42189c/frameset.htm

Regards,

Venkatesh

[Former Member · Accepted solution]

Hi Venkat

The link which you have given is good.

Thanks for the info! But i am not able to find anything about how and why semantic groups in TRFN is required.

Hi Suhas,

Thanks for the info! Can you please tell me where this setting for semantic groups in TRFN is available.

[former_member185132 · Accepted solution]

Hi,

In the TRFN change mode, the Semantic Groups option can be found on the toolbar. It's the third button from the right (between the End Routine and Currency Conversion buttons). Or you could find it in the Edit menu.

Regards,

Suhas

[Former Member · Accepted solution]

Hi Potter ,

Lets consider one scenario.

1. You have a 0WBS element coming from source and it is in your semantic key of DTP .So at the time of extraction all Data get sorted and packaged based on WBS element .

2. Now at your routine you are doing a lookup on network activity .So w.r.t. 1 WBS_ELEMT let say you can have 50 network activity (1:n relationship ).Now It means the amount of data will grow in each package which is sometime a performance overhead and we have nothing to change this except reducing package size but that too is not very helpful with semantic key as all records with 1 WBS_ELEMT should come in one package ,so we are totally dependent on data .

3. Now if you will have option to set semantic key based on network activity then it is an interesting thing because you may want to write some routine where all data should be sorted based on wbs elemt and network activity.

4. So its a kind of capability that SAP providing to handle data more efficiently .as data records with the same values in the selected fields thus form a logical key and are processed together in a package.

Hope this will be helpful .

Regards,

Jaya Tiwari

[esjewett]

There are a lot of good answers here, but I just wanted to point out something that has not been mentioned. There are two primary reasons for defining semantic grouping of packages:

1. Transformation logic (usually in the start or end routine) that requires access to all records with a particular key. For this type of logic, you need to have the semantic grouping applied all the time, so it makes sense to define it in the transformation itself. This was not possible before BW 7.3. You had to be really careful to always define the semantic grouping in new DTPs.

2. Error handling. The semantic grouping in the DTP defines the key of the error stack. If a record has an error, then all subsequent records with the same key are put into the error stack and not updated in the target, regardless of whether or not they have errors. This makes it possible to ensure that records don't get out of order when using the error stack. Since not every DTP necessarily uses error handling and we could conceivably want different semantics and granularity of error stack keys for different DTPs with the same transformation, it makes sense to define this kind of semantic grouping in the DTP. More here: [http://help.sap.com/saphelp_nw73/helpdata/en/42/fbd598481e1a61e10000000a422035/frameset.htm]

So, as you can see, there were really two different use-cases for semantic grouping. Use-case #1 was probably the most commonly discussed use-case, but it was not well-served by the BW functionality until the 7.3 release.

Cheers,

Ethan

[Former Member]

hi

Thanks a lot ppl! This solved my doubt.

[former_member185132]

Hi,

You're right, the setting can be done in DTP as well. As per the help, however, it is now recommended to do it in the transformation. My understanding is that it's because sometimes, the transformation may require that the data coming into it must be grouped together based on some characteristics. Before 7.3, we used to handle it in the DTPs by specifying the semantic groups there. But in that scenario, you might have a case where one DTP defines semantic groups as fields A, B and C, whereas another DTP (maybe due to human error) doesn't have a semantic group at all. This might cause problems when the TRFN processes the data.

Doing the semantic grouping in the TRFN however, will ensure that all DTPs of that TRFN will have the same semantic grouping that is defined in the TRFN. In addition, you can define a semantic grouping in the DTP as well - the effect will finally be that the semantic grouping will be the combination of the group specified in the TRFN and in the DTP.

Regards,

Suhas