Mapping Techniques determines Runtime Performance

[Former Member]

What mapping technique offers the best performance at runtime (i.e. graphical, ABAP, Java, XSLT)? I have the ability to use anyone of these techniques, but I want to ensure I am using the one that offers the best performance at runtime, not the quickest/easiest to implement at design time.

[Former Member]

Hi,

In the order of performance

1. Graphical (ease of development and fast)

2. ABAP Mapping

3. Custom Java Mapping

4. XSLT (worst of all): as it does an in memory transformation, if the messages are huge, it will pound performance.

Cheers,

Naveen

[Former Member]

Hi Naveen , i was under the impression that the performance should be almost same for Graphical mapping & custom java mapping , cos both use Java RunTime. If we create mappings using the graphical tool , XI generates a .class file in the background rite.

It ll help , if you can explain more on how you arrived at the performance figures.

Thanks

Saravana

[Former Member]

Hi..

imo mapping performance does only partially depend on technique used, another important influence is the developer himself. You can develop highly performant graphical mappings, java mappings and also XSLT mappings but you have to focus on performance, know the concepts and maybe have someone else look at the mapping after it is finished I'm not sure about ABAP mapping, but i would say graphical and custom java have the same performance. XSLT is indeed slower, but there are some problems which can not be solved in graphical mapping (e.g. recursion) which can be solved rather easily in XSLT. Also SAP uses XSLT extensively in their standard content (e.g. MDM 3.0) where we have not encountered serious performance problems so far.

best regards

Christine