15 MB text file

[Former Member]

Hi,

We have a 13 MB text file which we pick up using XI using content conversion.

But when the file is converted into XML, understandably the size of the whole payload becomes 6 times. Hence a huge size.

XI is not able to process this file at runtime after picking it.

What can we do to make XI to process this file?

The client will not split the source file. That is ruled out.

One method which I can think of is to use the content conversion parameter 'recordsets per message'. Suppose if the file has 100 recordsets, we can pick 10 at a time.

Kindly let me know all the possible solutions.

Thanks,

John

[abhay_rajhans2]

Hi John,

13 MB file is not a big file. In our scenario we have processed files of 60 MB size. And when a message in the processing its size appears to be larger but once it is processed then again size will be less.

Try to identify why the file is failing in XI. There might be UDF or other code or functions which are used that might be taking time.

In our case we had same type of problem the issue was with the SAP code and function if. SAP suggested to use ifs instead of if. In your case you can report the problem to SAP also but before that just check with your message mapping.

[stefan_grube]

13 MB is not much and should work, you have to increase memory space in PI Java settings.

For content conversion use element names as short as possible. 5 to 6 characters should be enogh to identify the elements in mapping.

If possible split the file in sender adapter, but this is not possible when the message has to reach the target at once.

Instead using contetn conversion, you can create a Java mapping to work on the text document. In this case you can omit any empty field.

[Former Member]

Hi Stefan,

How do we increase the memory/size in java stack (which you suggested)?

Also are there other setting parameter which can have an impact?

Thanks,

John