What are the options for unstructured Data supportability with TEIID?

[aswathkhan]

I'm looking for the following

   - How do we use Unstructured Data source (free flowing text like email contents) with Teiid?

   - Is there a possibility of integrating a Data Quality tool with Teiid?, If yes, let me have the details.

Thanks,

Aswath

[rareddy]

Aswath,

   - How do we use Unstructured Data source (free flowing text like email contents) with Teiid?

Teiid provides a public API to write translators for any source, see Developer's Guide and sample translators in the Teiid Code.  Using the translator concept you can map any kind of source into Teiid engine. However, how you map and what is target metadata for this all depends upon your interpretation of the data.

   - Is there a possibility of integrating a Data Quality tool with Teiid?, If yes, let me have the details.

Yes, you could do that in view layers. Possibly use the User Defined Functions for this purpose? I have not particularly embedded any external library. If you happen to know any good open source libraries in this area let us know, may be we can add a feature to integrate that.

Thanks

Ramesh..

[kenj]

Ramesh Reddy wrote:

Aswath,

   - How do we use Unstructured Data source (free flowing text like email contents) with Teiid?

Teiid provides a public API to write translators for any source, see Developer's Guide and sample translators in the Teiid Code.  Using the translator concept you can map any kind of source into Teiid engine. However, how you map and what is target metadata for this all depends upon your interpretation of the data.

Hi Aswath,

Following up on Ramesh's points, what is your use case for unstructured data as a source?  In other words, what do you hope to return from a query against an unstructured source?  Are you querying for a document that contains a string where you want the entire doc returned?  Are you looking to return metadata about the unstructured source such as author or number of words?  Is the unstructured content stored in some kind of system/application like a content management system?

-Ken

[aswathkhan]

Thanks Ramesh for the response. Let me start working on the same and will buzz you for any questions as I always do :-)

Ken,

   Thanks for the response.

    We have variety of use cases.

                 1) We do have repositry (shared drive) of our markieting files where we,

                       a)Pattern search and look for no.of occurnances of certain set of string(s) and try to align those words with our org data.

                       b) Also looking for the documents that matched certain string(s)

                 2) We do pull data from the certain internet sites do the string matches.

We do not have a content management system as such, but traditional shared drives where we have our email archives & documents.

Thanks,

Aswath