2

Short version of question - Does the index crawl through the content of uploaded documents, and if so, can that be turned off?

Longer version - I have a site consisting entirely of uploaded PDFs of various reports. Several custom columns on each report which are used to search for the report. Thee's no need to search the body of the reports for anything, only searching by name, user ID, run date etc.

I suspect the index searches the body of the docs, and as such the index crawls are taking a VERY long time - 4 DAYS for the full, 15 hours for the incremental.

If I can tell search not to crawl the documents themselves, I think it'll improve by an order of magnitude.

But this may beg another question - If I'm only interested in searching the metadata of the reports, is using the built-in search still the best and fastest method? Might I be better off using one of the other database query methods? It seems that Search would still be the best (assuming I can cut off crawling the documents) but I want to make sure.

1 Answer 1

0

Yes it will index the contents of documents but it depends.

Before moving forward - Query the database directly is a no-no, so I would take that away as an option. If all the documents are in a single library you can use the filters and views to create different access points to documents.

In 2010 - PDF document indexing is not included out of the box. If it has been enabled on your server there is a good chance it is using the Adobe iFilter - which is notorious for creating long crawl times.

You can tell search not to search PDF documents by doing the following:

  1. In Central Administration access the search service and look under approved file types. If PDF's are listed you can remove them.
  2. Check your Registry ( HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office Server\14.0\Search\Setup\ContentIndexCommon\Filters\Extension) and remove any PDF references. The extensions will be listed with GUID's so you have to browse thru them to see if you have Adobe iFilter loaded. The GUID is different depending on the version loaded.

NOTE: This affects your search service in it's entirety.

You can actually change some other settings - that I don't have handy such as timeout and document retry limits so Search doesn't spend too much time on any one document. I would also recommend checking out SQL during your crawls and see if you are getting any timeouts. If so- then improving the response time of your SQL server will go a long way to improving your Search performance.

4
  • I didn't mean hitting the database DIRECTLY, I meant using a CAML query or one of the other tools in SP. But if I can keep with search it's clearly the best scenario. Commented Dec 22, 2014 at 20:09
  • Hm, it seems to be skipping the PDF entirely now. I may not have been clear - I need the metadata of the document to be searchable, as in I need the document to be findable by entering the run date, provider ID, various custom columns connected to it, but I don't need the contents of the PDF indexed. Commented Dec 23, 2014 at 15:07
  • If you had the iFilter and disabled it - you can add back the PDF file type to search. It should then pick up the document properties (from SharePoint). The document itself won't be searchable though - when you search you will see that the URL reference will return the document info page (aspx) not the document itself. Commented Dec 23, 2014 at 15:13
  • Yep, figured that out between my last comment and yours. Looks good now, thx. Commented Dec 23, 2014 at 16:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.