Realistic Category Quantity

[thegman]

Hi

On a different thread I stated that I thought I had solved my problem by using some search modding & a vacant field.
OK. I was wrong. Zoom seemed unable to accept the mixed boolean logic required for what I needed to achieve.
(query = MUST CONTAIN "XXXX", MUST CONTAIN "YYYY", CAN CONTAIN "ZZZZ", CAN CONTAIN "111*", CAN CONTAIN "22**")
You get the idea.
If I am wrong & Zoom can handle such queries, without Categories, please let me know, as I can't get it to do it.

Assuming that it can't, does anybody have any realistic idea on how zoom handles a SEROIOUS quatity of Categories?

Let's say for 350 Categories, 10k pages. (Of which indexed content is next to nothing!!)(Using Meta Tag Method, not URL)

Thanks!!!

[Ray]

I thought that in your original thread you were using categories for the "must contain" part of your query, while the actual query would be the "can contain" part. But I must admit to being a bit confused as to what you are trying to do from the information you have provided.

No, you can not have a keyword query which mixes boolean AND and boolean OR request (it can be either, "match all search words" or "match any search words"). Not without custom development on the existing script anyway.

There is no explicit limit on the number of categories you can use. We have tested over a hundred.

Performance wise, it would seem that your approach is not the most efficient way of doing this (that is, using Zoom to search pages where you do not have any actual indexable content, and you merely want to recognize a predefined list of pages for given keywords). It would seem like a situation where a database would make more sense.

[thegman]

Ray
Thanks for coming back to me.

My application would be prodominently providing navigation.
Static HTML Product Pages are generated from a local Excel database.
(This database is VBA managed locally including various FTP supplier feeds. This method offers me far more control over the management of page content and FTP import filtering.)

My current server package does not allow for mySQL, I could upgrade, but looking at Database software offerings I am finding it dificult to visualise how I would integrate their generated forms. (So I may need to talk to someone who is familliar with their use.)

Where you say "without custom development on the existing script". in response to Mixed Boolean, are you referring to the search.php script, or the scripting of the engine itself?
As I have tried to get my head round if the Search.php could be modded to filter returned results based on parsed formfields.
(Bear in mind I could easily programitaly maintain the required parameters in a Search.php via VBA if necessary.)

I note your comment on performance & efficiency, but, when you tested on over a hundred, did it present major lag?

[Ray]

Where you say "without custom development on the existing script". in response to Mixed Boolean, are you referring to the search.php script, or the scripting of the engine itself?

I am referring to the "search.php" script. I don't think it would be a trivial change though and it would require an experienced developer.

I note your comment on performance & efficiency, but, when you tested on over a hundred, did it present major lag?

The search engine is designed to search indexed content quickly (this is the purpose of an index). What you are essentially asking for is the performance of searching for things which are not indexed. In our testing, we would always have actual content to index and search for, and the categories would only be an additional filtering parameter, so the results won't really correspond to your situation. Searching by non-indexed data will always be less than ideal.

[thegman]

Ray
Thanks for your reply.

Re PHP Modding. Understood. At least I know that it's future possible, which is reassuring. (Even if I might struggle myself.)

Re the Speed of filtering.
Surely I can draw a comparison, as my scenario is using some indexing, but just a lot less than yours. So Layman's logic would say that if your test was OK, mine should be.
I am of course indexing Meta's like Title, Keywords & Description. I am also indexing the Body, but all indexed body content contains nothing but a sitewide static ID key.
So that a filter with a zoomquery of that key would return ALL pages that pass the filter.
Product description, name, partcode, manufacturer etc. are all within the Description, keywords & title somewhere, so products can be properly searched.
I would be using the filters to aid navigation & effectively provide hiarachy/drilldown to the client.
So, while I appreciate that zoom is geared to searching indexed content, rather than non indexed filters, when you tested with indexed content & over a hundred categories, did it lag much?
(Currently I would only have 21 Categories for about 900 products, But as I increase the product quantities later, this number would also increase at a similar percentage. Thus my reason for asking is more to do with how much I would be able to increase by before addressing the search.php issue to resolve this)

[Ray]

I am of course indexing Meta's like Title, Keywords & Description. I am also indexing the Body, but all indexed body content contains nothing but a sitewide static ID key.
So that a filter with a zoomquery of that key would return ALL pages that pass the filter.
Product description, name, partcode, manufacturer etc. are all within the Description, keywords & title somewhere, so products can be properly searched.
I would be using the filters to aid navigation & effectively provide hiarachy/drilldown to the client.
So, while I appreciate that zoom is geared to searching indexed content, rather than non indexed filters, when you tested with indexed content & over a hundred categories, did it lag much?

No, but then I wasn't searching by multiple categories. That is, I would be searching for multiple keywords in the index, with only one category applied. While it sounds like you might be searching for very few unique words in the index, and depending on multiple categories to filter the results. Basically, the factors prevent me from making an estimation and giving you any sort of figure. You would have to test it on your specific scenario. There's always the CGI option if the PHP is slow though.

[thegman]

Ray
I would be searching by single categories initially anyway, as the quantity of categories I am making available are a result of offering maximum refinement of filter options.
I am sure that you understand that before purchase, running limited eval versions, we have limitations on testing scenario quantities. (Chicken & egg) Hence my question.
Which I feel that you have answered.
I appreciate that you need to be sensible about answers of this nature.
Frankly that endorses confidence, in that your approach is one of substance, rather than sales spin.
Thank you. It is appreciated, & very much understood.
(I have had dealings with an Australian Touchscreen Winamp Plugin called TouchTone, & I must say that their approach is similar. Maybe that is why the UK & US are in such a mess!! This conscientious approach to business in the UK is a real minority, trust me!!)
Thank you again.