Bug: re-Search not working when ? included

See original thread regarding the fix now in beta 2.

I compiled some Basic searches in Beta 1 and compared with Beta 2 (Match case and Match all word forms not selected):-

Query

Beta 1

(hits / resources)

Beta 2

 (hits / resources)

it 2, 510, 313 / 1087 2, 510, 313 / 1087
it? 467, 181 / 1081 2, 967, 473 / 1091
it NOTEQUALS it zero zero
it? NOTEQUALS it 457, 172 / 1081 457, 172 / 1081
it* NOTEQUALS it 641, 055 / 1087 641, 055 / 1087

My understanding is that "it?" in B2 should return the same results as "it" as the ? was being ignored. A quick look shows that it returns results for "its" e.g. The Canopied Earth: The World That Was.I'd also like an explanation for the 457, 172 NOTEQUALS results vs 467, 181 in B1.

Dave
===

Windows 11 & Android 13

Comments

  • Bradley Grainger (Logos)
    Bradley Grainger (Logos) Administrator, Logos Employee Posts: 12,190

    My understanding is that "it?" in B2 should return the same results as "it" as the ? was being ignored.

    ? is still a wildcard character. The problem is that the query "it?" is ambiguous: did the user really mean to search for all three-letter words that start with "it" (unlikely), or did they copy and paste the text in from some other source, or try to ask a question of the search engine (likely). We've enabled the latter scenario without disabling the former.

    I'd also like an explanation for the 457, 172 NOTEQUALS results vs 467, 181 in B1.

    I'm not at a Beta 2 computer right now, so I'm unable to repeat the experiment. What are your Match case and Match all word forms settings?

  • Dave Hooton
    Dave Hooton MVP Posts: 36,339

    What are your Match case and Match all word forms settings?

    Sorry, both are not selected.

    Dave
    ===

    Windows 11 & Android 13

  • Dave Hooton
    Dave Hooton MVP Posts: 36,339

    ? is still a wildcard character. The problem is that the query "it?" is ambiguous: did the user really mean to search for all three-letter words that start with "it" (unlikely), or did they copy and paste the text in from some other source, or try to ask a question of the search engine (likely). We've enabled the latter scenario without disabling the former.

    So "it?" now becomes a search for "it" or three letter words beginning "it".

    I found that the discrepancies above arise from text in footnotes (I had been searching "All Text"), whether in 5.1b or 5.2. Multiple results in the same footnote are counted only once, so the combined results from two queries can be larger than the result from an equivalent single query  e.g. in v5.1b  it? OR it won't give the same result as the results from each of the two terms in separate queries.

    If you search on Surface Text, though, things come together. In 5.2 beta:-

    it? NOTEQUALS  it  => 431,242  (equivalent of it? in 5.1b)

    it                           =>  2259095

    TOTAL                         2690337

    it?                        ==> 2690337

    I'm not sure if it? will now help people or confuse even further. But it's more accurate to say that ? is a wildcard character except when it can be mistaken for a question[:)]

    Dave
    ===

    Windows 11 & Android 13