Hi,

On 12/12/15 08:24, Feed My Lambs Esq. wrote:
I believe I found something quite different, though: A fundamental problem with full-text indexing where part of the XML schema gets placed inside the fulltext attribute (I believe this has just been reported in JIRA here: https://jira.duraspace.org/browse/DS-2948 about 6 hours ago).


Also, and more troubling, only the last bitstream ends up getting made available in the index for full-text searching.

This second one is separate from the Jira issue you linked to. Would you mind opening a new Jira ticket for the bit in boldface? Please include your observations made on the test server (the content of the test server re-sets over the weekend, but steps to reproduce the issue are always good to include).

It sounds very plausible, and not like what should be happening.

If there are two or more bitstreams for an item, only the last one tends to be available in the full text index. If anyone known a good person to contact for fixing a full-text indexing bug, feel free to forward this on to them, along with the examples.

I'm sure you're aware the DSpace is developed by a community of volunteers. Filing the issue in Jira will give it more visibility than posting it on the mailing list. Hopefully a volunteer can then be found to look into this.

We are leaning towards not using version 5.5 (unfortunately staying with version 3) if this particular full-text searching unpredictability might be semi-permanent. We have liked everything else we've seen about version 5, though, and hope to hear back soon about the nature of this bug and hopes for resolution (or workaround).

I'd imagine that if the code really does only index the first .pdf.txt bitstream, then this behaviour would have been in place since the Discovery-based search was first released. Unless you're using the old Lucene-based search in your version 3, I'd think there is no benefit in staying on version 3 over upgrading to 5.4 when it comes to search.

As a workaround, you could look into re-ordering the .pdf.txt bitstreams so that the most important one (if there is one) comes first.

cheers,
Andrea

-- 
Dr Andrea Schweer
IRR Technical Specialist, ITS Information Systems
The University of Waikato, Hamilton, New Zealand
+64-7-837 9120


--
You received this message because you are subscribed to the Google Groups "DSpace Technical Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/dspace-tech.
For more options, visit https://groups.google.com/d/optout.

Reply via email to