PassMark Logo
Home » Forum

Announcement

Collapse
No announcement yet.

Searching in Custom Meta Search Fields

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

  • Searching in Custom Meta Search Fields

    Using: Zoom Search Engine, v. 7.1 (Build: 1000)
    Indexing mode: Offline
    Platform: Javascript

    The files indexed are all pdfs. We are using .desc files because we want to be able to search by field.

    I have the Scan Options > .pdf extension configured to "Use description (.desc) files".

    My Custom Meta Fields are defined as:

    Meta name - Type - Search form name - Search Method - Values (dropdown only)
    TITLE2 - Text - Title - Partial text match - N/A
    PAPER - Text - Paper - Partial text match - N/A
    AUT - Text - Author - Partial text match - N/A
    AFF - Text - Affiliation - Partial text match - N/A

    A typical .desc file looks like this:

    <title>Pipeline Transport System</title>
    <meta name="title2" content="Pipeline Transport System">
    <meta name="paper" content="ABC2015-1234">
    <meta name="aut" content="Brad Adams, Christina Barnes">
    <meta name="aff" content="Open University">

    I have the fielded search boxes on my search page for Title, Paper, Author, and Affiliation, along with the default full text search field.

    I also have the ZoomShowMatchOptions() options text enabled:
    Match: _ any search words _ all search words

    If I search for "pipeline", "pipeline transport", or "pipeline transport system" in the title2 custom meta field, I get the above hit, regardless of whether "any" or "all" search words is selected.

    If I search for "pipeline system" in the title2 custom meta field, I do not get any hits, again regardless of the match options. (This example searches for words in the title that are not consecutive.)

    Likewise, if I search for "transport pipeline" in the title2 custom meta field, I do not get any hits, again regardless of the match options. (This example searches for words in the title in a random order vs the order in the title.)

    The other custom meta search fields seem to have the same issue (i.e., you have to search for phrases exactly as they appear in the meta data).

    Is there a way to get these types searches to work with the Javascript platform?

  • #2
    The custom meta field feature TEXT fields are not designed for tokenized/single-word-indexed searching ala the <title> field or meta keywords field, or any other content on the page.

    It is more for single word or exact phrase values (e.g. names, locations).

    To explain it in more technical detail (should you be familiar with programming/scripting) -- the <title> field is indexed as part of the page, and the individual words are added to the dictionary. This is why you can search for it in the main search box, and any individual word in any order will yield results.

    But the title2 meta field you have specified is stored separately from that, as an entire string. The query "pipeline system" can only be compared to "pipeline transport system" via either a strcmp (exact match) or a strstr (substring match). This is why it is not matching.

    To change this would require a significant overhaul of the custom meta field feature, so it is not something we would be considering at this point unless there is alot of user demand.
    --Ray
    Wrensoft Web Software
    Sydney, Australia
    Zoom Search Engine

    Comment

    Working...
    X