Removing 'the' from search results and front end filter question

[Chris Roberts]

Hi guys– Thanks for always answering my annoying questions and for always being super responsive and non-judgemental ;-).

Our client has a gazillion articles dating back to 2003 or maybe earlier so it’s a process getting the search configurations sorted. At the moment I’m using the index table engine as it was recommended in the setup guide and we have had the best results with it so far. However, I notice that we cannot use a lot of front end generic filters which might be useful for more exact results the first time. Why are they not recommended? just curious so I can explain to the client. They would like to be able to sort the results alternatively by selection between relevance and date. I noticed the date span setting but is there a sorting option? So, check clickable radio buttons in the front end search box for either Relevance or Date. Is that possible?

Also at the moment we are getting ‘the’ highlighted early and a lot in results. If we type in, for instance, ‘The Other Gulf War’ (an article from 2005) in the search box, it doesn’t come up as a match until page 5 (notice all of the ‘the’ words highlighted) although it should be an exact match to the title. The ‘exact match’ thing didn’t come up until we turned a lot dials up for chronology to satisfy the client’s desire for ascending order. Then, Authorship became a priority so we turned some more dials to get better results for guest authors (you may remember this discussion or ticket) and so far so good but now we’ve lost the ability to type in the title of an article and see it come up at the top. That seems like it should be pretty obvious and probably would work nicely if we hadn’t already screwed around with the other settings.

I have searched the support forum and general Google search for this issue but haven’t seen anything helpful yet. Thanks for any help or recommendations, guys.

-Chris

[Chris Roberts]

Also, I forgot to mention that I did add ‘the’ and a few other words to the STOP settings but it doesn’t seem to be helping.

[Ernest Marcinko]

Hi Chris,

Thank you very much for your kind words, I do try my best to answer and help with everything I can.

The reason for not recommending filters is purely for better pefromance. The index table uses so called indexed queries, which is a database feature, but there are certain rules that a query needs to obey for the database engine to use the index and not to switch to different methods automatically.
This is the main reason the index table can be significantly faster and less resource heavy. Adding filters to the mix is going to reduce the chances significantly of having indexed queries and the database will use different methods more often – because with filters table joins must be used and joining a huge set of data with another to get the connections between the objects can be resource heavy.

Still, you can try them, it very heavily depends on the datbase and the exact scenario. If you see significant slowdown, then it is probably because of that. You can also the index plugin to make more indices on some of the core tables, that can help greatly.

As for the second issue, I need to check, however I can’t login with the details. Can you please check the username and the password?

[Chris Roberts]

You cannot access this content.

[Ernest Marcinko]

For now I think WP Admin access should be sufficient. I only tried that, but for some reason it does not let me in. Can you please try making a new user? Thank you very much!

[Chris Roberts]

You cannot access this content.

[Ernest Marcinko]

Thank you, that worked!

I think I found the problem, your primary ordering is set by “date descending”, make sure to change them up like this: https://i.imgur.com/db1ehVu.png
.. so relevance is primary and whatever is secondary. That will do the trick 🙂

[Author]

Posts