Which work item attributes are included in full text search?
We are wondering which attributes are included in full text search. From the below, it seems that Large String type and Large HTML type are included, and we suspect that many other text type attributes are included.
Customization of Work Items in Rational Team Concert
https://jazz.net/library/article/1003
Our understanding is that full text search looks at the JFS text index in the jfs-textindex folder, so thought we might have the answer by looking at the files in there, but we couldn't interpret that. We are under the impression that non-text type attributes, such as enumerations are not included in full text search.The articles we found talk about how full text search will split up the search terms using a tokenizer, but they don't seem to talk about which fields are included in the search. Could someone explain?
Workitem Quick Search in IBM Rational Team Concert
http://www.ibm.com/support/docview.wss?uid=swg21586008
Getting the most out of full text search
https://jazz.net/library/article/824
We are talking about the Full Text search you get from the Full Text query condition, and from the Quick Search field in top right hand side of Web UI window. We assume the search provided by these two is identical and searches on the same attributes. Please correct us if we are wrong.
Note: We read this article, because we thought that part of the answer lay in understanding how the JFS text index is created in the first place. (If that doesn't contain certain fields from work items, then of course the full text search will not be searching on those fields.)
Query and Search indices management in the Rational solution for
Collaborative Lifecycle Management
Part 1: Query, Search and indexing technologies in CLM
https://jazz.net/library/article/1271
We would appreciate any clarification on this.
Customization of Work Items in Rational Team Concert
https://jazz.net/library/article/1003
Our understanding is that full text search looks at the JFS text index in the jfs-textindex folder, so thought we might have the answer by looking at the files in there, but we couldn't interpret that. We are under the impression that non-text type attributes, such as enumerations are not included in full text search.The articles we found talk about how full text search will split up the search terms using a tokenizer, but they don't seem to talk about which fields are included in the search. Could someone explain?
Workitem Quick Search in IBM Rational Team Concert
http://www.ibm.com/support/docview.wss?uid=swg21586008
Getting the most out of full text search
https://jazz.net/library/article/824
We are talking about the Full Text search you get from the Full Text query condition, and from the Quick Search field in top right hand side of Web UI window. We assume the search provided by these two is identical and searches on the same attributes. Please correct us if we are wrong.
Note: We read this article, because we thought that part of the answer lay in understanding how the JFS text index is created in the first place. (If that doesn't contain certain fields from work items, then of course the full text search will not be searching on those fields.)
Query and Search indices management in the Rational solution for
Collaborative Lifecycle Management
Part 1: Query, Search and indexing technologies in CLM
https://jazz.net/library/article/1271
We would appreciate any clarification on this.
One answer
Full text search in CLM is powered by Apache Lucene.
https://lucene.apache.org/core/
The full text search content folder for RTC is usually server/conf/ccm/indices/workitemindex/fulltext_index. You can examine the content in some third-party tools, such as Luke.
https://code.google.com/p/luke/
https://lucene.apache.org/core/
The full text search content folder for RTC is usually server/conf/ccm/indices/workitemindex/fulltext_index. You can examine the content in some third-party tools, such as Luke.
https://code.google.com/p/luke/
Comments
Thanks for the tip on Luke. I am kind of hoping not to have to use Luke to debug the index file to figure out what work item attributes it contains. We are looking for a more general answer. Is there someone from Foundation who knows this?
You don't use Luke to "debug" the index. You "interact" with the index - create, delete and search documents. As for what attributes are indexed, if I read the article #824 correctly, all texts are indexed, regardless where the text appears. The article specifically states that even attachments are indexed. If I'm not mistaken, the literal value of an enumeration is also indexed.
Do you find any text that is not indexed?