Are there any plans to include exact matches within the index table?

[pbaines]

Hi Earnest,

Thanks for your help and support on other tickets. I have a quick question regarding the exact matches within the Index table setup, I can currently see that its not possible to search for an exact match post title e.g. ‘test post title name’ and the post ‘test post title name’ will appear in the results, when using the index table.

e.g. https://ispnlibrary.org/article/comprehensive-analysis-of-hypermutation-in-human-cancer/ is not found when searching for ‘
Comprehensive Analysis of Hypermutation in Human Cancer’, as you can see here: https://ispnlibrary.org/search/?asp_s=Comprehensive+Analysis+of+Hypermutation+in+Human+Cancer&p_asid=1&p_asp_data=1&aspf%5Bpost_date_from__1%5D=&customset%5B%5D=external-resource&customset%5B%5D=meeting&customset%5B%5D=video&customset%5B%5D=presentation&customset%5B%5D=article&customset%5B%5D=page&post_order=&filters_initial=1&filters_changed=0&qtranslate_lang=0&current_page_id=622&asp_s=Comprehensive%20Analysis%20of%20Hypermutation%20in%20Human%20Cancer

However, is there or will there ever be a way to use the performance of the index table and have exact matches? Currently, we are making a decision between faster results loading, but less accurate OR slower loading results, but more accurate (based on exact match searches)

Thanks,

Peter

[Ernest Marcinko]

Hi Peter,

It’s unfortunately not possible to do so in general. Index table is basically a list of keywords, where every keyword is in a separate line (sort of like a spreadsheet). Therefore the order of the words is lost and so exact matching is impossible.

However the post titles are an exception as they are short. They are stored in two ways – both words and the whole titles up to 50 characters in the keyword field (that is the keyword length limit).
Basically if the title is selected for search an exact and a partial query is also tested against these titles.

In your case it may not match for 3 possible reasons:
– The title is longer than 50 characters (which is true in this case)
– Also, if there are stop-words enabled and the stop word is detected during the index then the exact title is not indexed. This could be true as well, as if you have that enabled “of” is in the title any by default it’s excluded.

However the results should still show up, so I’m guessing that you are using the “AND” or “AND with exact keyword matches logic”, and I’m guessing that you have stop words enabled – so the “of” word in the title is casusing the non-match. So I would suggest two possible solutions:
1. Either disable the stop-words and re-index
2. Or add the same words on the search instance keyword exceptions list as well, so they are removed from the search phrase

Either solution should do the trick.

[pbaines]

Brilliant, thanks! Just adjusted the search instance keyword exceptions list and lts now showing results, even for longer titles.

[Ernest Marcinko]

Great 🙂

If you don’t mind, I will close this topic soon and mark it as resolved, feel free to open another one if you have other questions or issues.

[Author]

Posts