How to implement a broad search on the title field in DSpace 7/8?

[Lucas Silveira]

How to implement a broad search on the title field in DSpace 7/8?

Hello everyone!

I'm trying to implement a broader search in DSpace 8 using the /discover/search/objects endpoint to filter by title (f.title). Currently, I'm using the contains operator, but I've noticed that it does not return all the expected results.

Problem

• When I search for f.title=Teoria geral dos partidos,contains, it correctly returns an item titled "Teoria geral dos partidos políticos".
• However, when I search for f.title=Teoria geral dos políticos,contains, no results are found, even though the string is very similar to an existing title.
• In other words, the search seems to work more like a partial equals than a true "contains" search within the title.

What I have tested

• Switched between /discover/search/objects and /discover/browses/title/items, but both have this limitation.
• Tried reordering words and removing accents, but no difference.
• Checked Solr, and the titles appear to be correctly indexed.

What do I need?

I want to implement a broad search on the title field so that users can find an item even if they don't type the exact title. Some possible solutions:

1. Is there a way to use wildcards (*) within contains?
2. Can I use another operator (fuzzy, wildcard, phrase, etc.) in DSpace searches?
3. If the solution is within Solr configurations, what adjustments should I make?

Any suggestions would be greatly appreciated! 

[Andrew Thompson]

You could try the extended dismax query parser.

It has a different query syntax. My organisation wanted this behaviour too. Config is in the schema.xml file I believe under the search core.

You might also be interested in partial contains so for example searching “Teori” matches documents with “Teoria”. I was able to achieve this by using an Edge NGram Tokenizer on the index filter chain. 

-Andrew

On 8 Mar 2025, at 6:13 am, Lucas Silveira <lucasangelo...@gmail.com> wrote:



--
All messages to this mailing list should adhere to the Code of Conduct: https://www.lyrasis.org/about/Pages/Code-of-Conduct.aspx
---
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 dspace-tech...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dspace-tech/6edd3557-24a5-4b39-be45-a90864c56b3an%40googlegroups.com.

[Lucas Silveira]

Hi Andrew,

Thank you for your suggestions! The eDisMax query parser and the Edge NGram Tokenizer approach sound very promising. I'll research both and test them in my setup.

I’ll get back to you soon with updates. Thanks again for your help!

Best regards,
Lucas