Today I wondered how to get #OpenText #BizMapper to insert a data row into a #database table, after processing the input source that is read from a #DB, while it constructs a document output to the file system (or an #mq, not a db).

It does not seem possible.

The solution is to turn around the process: get the data row entered first, then use that to construct the output document, then update the data row in post.

#ETL #dataMapping #EnterpriseMessagePump

Last week I migrated one of my #dataIntegration #software applications. It went from #XML input and #JSON output to JSON input and output, and at the same time from #Oracle #Java8 to #IBM #Java11.

That move cut away 2/3rds of the #sourceCode.

The functionality stayed roughly the same. Some validations got omitted. I may wind up adding them back in.

This is the way forward. It wouldn't have been possible without tooling that produces JSON documents. #OpenText #BizMapper
Did I mention #OpenText has an office on the other side of my town? I've become quite familiar with the #BizLink, #BizMapper, and #BizManager products they bought from #GXS. BizManager has some #usability and #accessibility issues. And its web interface is written in #JSP. I know #java, #webstandards, and JSP from long ago (BizManager's JSP seems that old). Maybe I should offer them my help?
Last week my team learned about a limitation in #XML processing in the #BizMapper application by #OpenText (formerly by #GXS):

If an XMLElement has 2 or more children with the same name, and the source format for that element has its sort order set to unordered, then BizMapper no longer knows which child element to pick.

We wanted the 2nd. BizMapper gave us the 1st.

This is understandable behavior. It's also quite annoying. We reached out to their support and hope for a solution.
The solution to this failure is 2-fold:
- recognize that the field got assigned a #namespace;
- set the field's namespace to empty (don't remove it!).

#businessIntelligence #BI #BizMapper #OpenText #XML
Working with #XML documents is a joy. Working with ones that have a #namespace is a royal pain.

The #BizMapper transformation tool for enterprise business communication by #OpenText (formerly by #GXS) can translate XML into #EDI: great for integration with our ERP.

If an XML document has a default ns, the Mapper happily adds it to newly-added fields. Normally that is great.

Until you don't expect it, and wanted no ns.

Then translation will fail.

#businessIntelligence #BI
Anyone else here who uses #BizMapper by #OpenText (formerly by #GXS)?

It's a document transformation tool for enterprise business communication (orders, acknowledgements, etc).

The latest version can read and produce #JSON: nice when we communicate with a partner who also speaks JSON.

It doesn't seem to want to use a JSON Schema as a template for source or target documents. That costs a lot of time.

Any ideas on whether it can do that? And if so, how?

#businessIntelligence #BI