topic Solr Does Not Index Xml Elements / Attributes in Alfresco Archive

Solr Does Not Index Xml Elements / Attributes

larophel — Fri, 23 Aug 2013 18:51:33 GMT

Hello,I have noticed that Solr sometimes does not index the names of Xml elements and Xml attribute values.Element content, on the other hand, is always indexed. I am using Alfresco 4.2.c Community, on Linux. The installation was performed using the default options.I put together a list of steps, wh

Re: Solr Does Not Index Xml Elements / Attributes

larophel — Fri, 23 Aug 2013 19:28:11 GMT

I forgot to mention: the Xml file must have the "<?xml version="1.0" encoding="UTF-8"?>" header, otherwise the problem cannot be reproduced.

Re: Solr Does Not Index Xml Elements / Attributes

larophel — Mon, 26 Aug 2013 14:54:20 GMT

I have found the cause of the problem.
Alfresco uses a "Transformer" to obtain the to-be-indexed text from a node.
For the mimetype "text/xml", there are two possible transformers: "TikaAuto" and "StringExtracter".

The "TikaAuto" transformer only returns Xml element text content.
The "StringExtracter" transformer returns the full Xml file content.

For each transformation, Alfresco measures the time it takes to perform the transformation.
Which transformer is chosen depends on the average time for past transformations.
So, this is why sometimes the full text content is indexed and sometimes only Xml element text content.