Only found 1st Occurence of word

[WilliamJ]

Hi again,

Zoom only seems to be finding the first occurence of a search word on a particular page at times. Other pages it seems to pick up more but not all. Have I done something wrong or do I need to do something in configuration that needs to be changed? These words that it is missing are not imbedding in any script or anything - normal HTML coding.

Thanks.

William

[WilliamJ]

1st Occurence

Figured this one out - The results page shows "TERMS MATCHED: 1" and I assumed it must only be finding the 1st occurence. I only had to increase the context from the default of 30 characters to show all the terms revealed. I wound up with 120 characters to display on the search results page.

[Ray]

The "Terms matched" value only indicates the number of search terms from the user query that has been matched. It does not indicate the number of occurances of the word on that page.

So if a user is searching for "dogs cats monkeys"

And a page contains 300 instances of dogs, and 50 instances of monkeys - it would say "Terms matched: 2".

The "Score:" value is more of an indication of the occurances of the words on that page. However, it is not a direct relation, as it is dependent on the weighting of the words based on whether they were in titles, boosted pages, etc.

Please see this page for a detailed description of our scoring system:
http://www.wrensoft.com/zoom/support/faq_score.html

[RobertS71]

All,

I have the exact same problem as William.

Only the first occurrence of a search term on a page is listed on the search results page.

I understand that "terms matched" and "score" have nothing to do with this problem, as explained by Ray.

SO... can someone please tell me exactly what to do to fix this? I need ALL occurrences of a search term on a page to be listed as an individual hit on the results page.

For example:

There are three manuals for the "GS7000 GainMaker Node" on the same page on our technical publications library web site. When searching on that term, Zoom Search only lists the first one on the page in the search results. Our users need to see that there are actually three manuals available for that product.

How do I fix this?

[David]

If you have 3 manuals then you should index the manuals themselves and not just the page that mentions them. Then you would get 4 results. 1 per manual + the summary page.

Zoom displays up to 3 pieces of context for any particular page. The number of search words determines the amount of context.

You don't really want to see all occurrences from a page, in the case of a common search word, this would result in the entire document being dumped out. Which is not good if you have large pages or PDF documents.

[RobertS71]

If you have 3 manuals then you should index the manuals themselves and not just the page that mentions them. Then you would get 4 results. 1 per manual + the summary page.

Ok. How do I index the manuals themselves? I am currently using spider mode for indexing. I do not index the internal content of any of these PDF manuals. I only index the html pages in our web site that they are listed on, i.e the pages that contain links to them.

Zoom displays up to 3 pieces of context for any particular page.

I don't quite understand what that means. Can you please give me an example of that?

Thanks for your help!

[Ray]

Ok. How do I index the manuals themselves? I am currently using spider mode for indexing. I do not index the internal content of any of these PDF manuals. I only index the html pages in our web site that they are listed on, i.e the pages that contain links to them.

You need to have PDF indexing enabled. This requires the PDF plugin. See the instructions on that page and in the Users Guide or FAQ section for information on indexing PDF files.

I don't quite understand what that means. Can you please give me an example of that?

One piece of context description is a segment of the words surrounding the keyword which was found on the page. For example, a search for "dog" may yield a single piece of context description like "... I took my dog out for a walk the other day and ...."

Zoom will display up to 3 segments of context at most for any single search result. It will not return every line/sentence that the word appeared on within that document. In most cases, this is almost never what you want, because it can be an unwieldy amount of information. Note that it does search and count every instance of the word (and it takes this into account for the relevance scoring) - it just won't be showing every instance on the search results page (similar, of course to all popular search engines like Google and Yahoo).

[RobertS71]

Raymond,

Thanks very much for the additional explanations.

I won't be indexing the contents of my PDF files. So I'm OK with that.

I understand now what you mean by context descriptions in the results page. I dont' think that I've ever seen more than one context description listed for a specific web page in my search results.

If I have more than 3 hits on a particular page, which I do, shouldn't that result in at least 3 individual context descriptions for that page in the search results? Or is there a setting I have to make somewhere to set how many I want?

[MergeThis]

The results description is simply the first nnn characters in the body of the topic. You set the "Max. description length [nnn] characters" in the "Limits" tab of the Zoom Indexer Configuration window.

That's the way it's designed, and that's the way it works, much like other search engines you've seen on the web. It's essentially a "preview" of what the topic contains, which users can view as a guide to whether they should click the link or not.


Good Luck,
Leon

[Ray]

I won't be indexing the contents of my PDF files. So I'm OK with that.

If you want to index the manuals themselves (as mentioned before), and wanted to have search results which link directly to each of the manuals (which is a much better result than just linking to the page containing the links to the manuals), then you would want to index the PDF files. However, it is up to you.

I understand now what you mean by context descriptions in the results page. I dont' think that I've ever seen more than one context description listed for a specific web page in my search results.

Most likely because you've only been searching for one search term, eg. "dog". There is a context description for each search term (not each match). So if you search for "dog cat", and there are 20 matches for dog and 3 matches for cat on the page, you will see 1 context description for the first occurance of "dog" and 1 context description for the first occurance of "cat". This gives you a total of 2 context descriptions (sorry if this is sounding like an episode of Sesame Street! )

As Leon mentioned above, you can increase the size of the context descriptions, but not the number of. There's a maximum of 3 per search result (much like Google, etc.).