TREX and file indexing/searching

[Former Member]

Hi out there!

Actually I have the task to index certain documents (PDF and TIF) which are saved via DMS in an archive. They should get indexed from TREX so that a later search is possible.

Problem: absolutely no competence with TREX engine So please be so kind and explain a bit in details...

Questions:

- How can I transfer/index documents to TREX? (ABAP or JAVA Client? code samples?)

- How does that indexing of documents work? Have I to provide key words or does TREX compute them out of the document content and document attributes? Or are both ways possible? If possible... code samples

- Some words about search performance

I hope you can help me asap

Kind regards

Kai

[KarstenH]

Hi Kai,

there is an age-old standard connection of TREX to PLM DMS described in PLM DMS' documentation. It addresses TREX vai an RFC connection. Thus, the first steps for this would be installing TREX and maintaining it as an RFC destination.

TREX produces a full text index. You do not have to provide the keywords for that. But - "TIF and PDF" - are these by any chance scanned texts? Because OCR is not included in TREX.. if your text is really in bitmaps, TREX will only index the attributes of the documents.

There is also newer ways to address TREX from PLM DMS, using the SES (search engine service for business objects). You can use SDN-search to indetify info on that. However, it does not change the situation with respect to OCR.

This is available for systems basing on NetWeaver 7.0 or higher. For the latest on that, please contact SAP's PLM consulting.

Own programming against TREX engine APIs is not documented and/or supported by SAP.

Best, Karsten