Content Management and Capture

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

Hi Jana,

this isn't by any chance a customer we both know :-)?

The only way that I can think of is to create a subclass of the class in question and make that full text indexed. When someone/somethings determines that full text is needed change the document class.

This - as you know - might have some side effects and you might have to update search templates and the like...

Let me know if you come up with something better.

Kind regards,

Gerold

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

I would say that it depends which search engine you are using. 

With Watson Explorer (old Content Analytics) it is possible to create custom crawler - but unfortunately WEX doesn't support content based retrieval (CBR), so it cannot be used in Filenet query.

I would bet that with Elasticsearch or OpenSearch it is possible, but you have to find expert from this domain (and customer has to have latest FN version - 5.5.12)

But with "default" Content Search services I think no such customization is supported, although Lucene (it is based on) may allow it.

Btw. if you have control over the Filenet QUERY (with CBR) than simply add where condition into it, filtering out those documents that you don't want to return (index) - if CRB is used in the query. Or response filter in ICN.

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

Hi all, 

thanks for the answers. For now, the client is satisfied with creating a separate subclass. As we are not going to implement it yet we will also look at Elastic search after the upgrade (to 5.5.12) planned for June in our case. 
But I am confused to version of CPE, I've read that "Content Platform Engine 5.5.8 introduces a technology preview of a new content-based search feature that uses either the Elasticsearch search engine or the OpenSearch search engine"
resource: https://www.ibm.com/support/pages/technology-preview-new-content-based-search-feature-content-platform-engine
It sounds like Elastic search can be used from 5.5.8. But also 5.5.12 is OK for us, as we plan to upgrade soon.

BR, Jana

I would say that it depends which search engine you are using. 

With Watson Explorer (old Content Analytics) it is possible to create custom crawler - but unfortunately WEX doesn't support content based retrieval (CBR), so it cannot be used in Filenet query.

I would bet that with Elasticsearch or OpenSearch it is possible, but you have to find expert from this domain (and customer has to have latest FN version - 5.5.12)

But with "default" Content Search services I think no such customization is supported, although Lucene (it is based on) may allow it.

Btw. if you have control over the Filenet QUERY (with CBR) than simply add where condition into it, filtering out those documents that you don't want to return (index) - if CRB is used in the query. Or response filter in ICN.

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana

I would say that it depends which search engine you are using. 

With Watson Explorer (old Content Analytics) it is possible to create custom crawler - but unfortunately WEX doesn't support content based retrieval (CBR), so it cannot be used in Filenet query.

I would bet that with Elasticsearch or OpenSearch it is possible, but you have to find expert from this domain (and customer has to have latest FN version - 5.5.12)

But with "default" Content Search services I think no such customization is supported, although Lucene (it is based on) may allow it.

Btw. if you have control over the Filenet QUERY (with CBR) than simply add where condition into it, filtering out those documents that you don't want to return (index) - if CRB is used in the query. Or response filter in ICN.

Hello all,

our client wants to use full-text search but the client wants to have control over what documents (of the same class) will be indexed. As far as I know, this is impossible, but I want to ask if anyone has any idea how to do it. How to index only some of the documents of the class. For example, I was thinking about cancelling the Class index job and creating only an Object Index job per document, but I don't know if this will work. Many thanks, Jana