1

SharePoint 2010 Enterprise Search.

We have a document library that has lots of PDF documents. Before PDF are uploaded to the document library they have property (metadata) called Title and has the same title since the PDF is used as a template. The title property has the following entry: Standard Design Template

When users upload the document they leave the SharePoint title empty and only change the name as shown below:

enter image description here

Based on the above screen shot, if I search for Manometers I see the search results showing Standard Design Template as the header as shown below:

enter image description here

If the PDF document has an empty Title (The title field here is PDF property field not the SharePoint) and uploaded to SharePoint document library and do a search for the name of the document we get the name of the PDF document in the search result as the header instead of the Title property of the PDF document as shown in the above screen shot.

I need to know how search determine which value to display in the search results. Also, if users enter a value in the Title field when uploading the PDF document to SharePoint, then search would display the value as the header instead of the PDF internal Title property.

1

Don't confuse list and library properties (your first image) with Managed Properties which are surfaced in search results. Managed properties are extracted by the search indexing process and in your case the Title property, the one embedded in the PDF, is being added to the index as the actual Title of the document.

You have a couple options and the one you choose depends on how painful the change will be for your organization.

  1. Change the Title in the PDF. This is the direct approach. The search service on the next crawl will pick up the change and fix the results.
  2. Change the Search Results page logic. The default page looks at the Managed Properties and returns Title. If title is blank it returns the Filename. You can detect the string "Design Standard Template" and render the Filename instead, but this only masks the issue.
    1. Populate the Library Title property for the PDFs. Next ensure that for the Title Managed Property you have the ows_Title crawled property above the crawled property from the PDF documents.
0

There is kind of a "preference" list in SharePoint Search Metadata settings that determines which fields to use in what order to present the title in the search results.

It can be found in: Central Administration>Manage Service Applications>(Your Search Service Application)>Metadata properties

There you look for the entity "Title" (it's usually found on page 3/3, as only 50 managed properties are shown in one page. Open the settings for the "Title" property.

There you find an option to assign this managed property to crawled properties: Crawled properties that are assigned to this managed property

SharePoint Search appears to determine which entity to display as title top-down from this list. By default, "Base:displaytitle(Text)" is on top of the list. This seems to be some kind of multi-level entity that tries to extract the title from the document (it also uses the first words in a PowerPoint presentation or the first heading in a Word document). If it cannot extract any of these, it uses the "Title" property from within the document.

For a more "unified" behavior, in the screenshot above I have added "ows_FileLeafRef(Text)" via "Add Mapping" and put it on top. "ows_FileLeafRef(Text)" is the filename of the document. So if SharePoint search can determine a filename, it will show that. Below is "ows_Title(Text)". That is the "Title" property of a SharePoint document or list entry. For Microsoft Office applications, this can also be edited from within Word, Excel etc. in the "File" menu. "E-Mail" and "PreferredName" may relate to list entries which are contacts (they were in the list by default).

My goal was to always show the filename, since my users were unsatisfied with the "Try to extract some text from the document" logic. Putting "ows_FileLeafRef" at the top of this list achieved exactly that.

Also I'm quite positive that all other entities in the list are still being crawled and thus are searchable - they only aren't presented in the search result page.

Please note: All settings only take effect after the documents have been crawled again, so you'd most likely initiate a full crawl after changing the metadata settings, so the crawled property settings get applied to all search results.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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