PassMark Logo
Home » Forum

Announcement

Collapse
No announcement yet.

Two join chars in a row

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Two join chars in a row

    I am using hyphen, apostrophe, comma and colon as join chars because we have chemical names that use these chars. From a thread I initiated a year ago, I had this response - **Two single-quote characters in a row will not be indexed, it will split the word. A "word join" character only allows it to join characters which are actual alphabet/numeric.**

    I have been trying to find a way to use Zoom with our site. I have the skip words length set to skip less than 1, because single char D and L are critical that we index, as well as single digits.

    My observation is this - I search for 1',2',3',4',5'-13C5 which is a distinct label pattern I would like to find. I don't get any results because (I guess) the 2 join chars in a row are splitting the terms. I search for "1 ',2 ',3 ',4 ',5 '-13C5" with spaces inserted before the apostrophes but not after, and the whole string surrounded by quotation marks and I get 22 results found, which is what I want.

    My question, is this expected behavior, when there are 2 join chars in a row? The terms are split, with the 2 join chars sticking to the front of the second term? I recently upgraded to 1028 and I had been using 1023. The site in question can be seen at www.omicronbio.com/search.php. Please note that I am entering these terms in the Search for box and leaving the product name box blank.

    Also, if this is the expected behavior, then what should happen when there are three join characters in a row? In order to find 2',2''-2H2 I found that searching for "2 ',2 '-2H2" works, but not sure why. Three join chars in a row means the first one is dropped and the second two stick to the front of the following term?

    The suggestion a year ago was to use a custom meta field, which I have incorporated, but I have problems with that because (I think) a search for Salt Sodium will not find a page with Sodium Salt in the custom field.

    Thank You

  • #2
    Originally posted by nmyers View Post
    My question, is this expected behavior, when there are 2 join chars in a row? The terms are split, with the 2 join chars sticking to the front of the second term?
    That's correct, it is the behaviour in V6, although it was not documented. It was originally to handle situations like C++ and C# but it does not make much sense for it to happen with characters preceding the word. So this is changed in V7. It will instead be indexed more like:

    1' 2' 3' 4' 5' 13C5

    We also need to make the search script (PHP, etc.) front-end behave the same way so that when you enter in "1',2',3'" it will actually only search for "1' 2' 3'".

    Originally posted by nmyers View Post
    Also, if this is the expected behavior, then what should happen when there are three join characters in a row? In order to find 2',2''-2H2 I found that searching for "2 ',2 '-2H2" works, but not sure why. Three join chars in a row means the first one is dropped and the second two stick to the front of the following term?
    Yes, that's an accurate explanation of what is happening in the current version.

    Originally posted by nmyers View Post
    The suggestion a year ago was to use a custom meta field, which I have incorporated, but I have problems with that because (I think) a search for Salt Sodium will not find a page with Sodium Salt in the custom field.
    You can have Partial Matching for a custom text meta field. This will allow "salt" to match "sodium salt" and "salt sodium".

    But yes it probably makes sense for the custom meta field to only be used for the chemical pattern and not for the text names.
    --Ray
    Wrensoft Web Software
    Sydney, Australia
    Zoom Search Engine

    Comment

    Working...
    X