Query Cache v1.0

[data_digger]

Hello,

This is my first official contribution.

I've created Query Cache contribution that might help improve performance of Zen Cart. It prevents Zen Cart from executing the same queries many times, which can speed up Your store. I tested it on few stores and it usually reduces number of queries from 30% to 60%. Installation instructions and contribution can be found here:
http://www.data-diggers.com/query-cache
There's also screencast showing how to install this module. Please let me know what You think about it. In case of any issues please contact me via pm or via email (data-diggers@data-diggers.com)

Note: there will be new version of this module, which will further reduce number of queries sent to MySQL.

[Marg]

Thank You!
Amazing results! From 1060 down to 472 on one of my sites...

[data_digger]

Happy to hear that:)
If You've found any bugs or have any suggestions please write and I'll try to fix bugs/implement new features.

[wastaboo]

Thank you i got a 60% nice job

[data_digger]

New version of Query Cache is available. In nutshell, it reduces query count by 80% ( v1.0 reduces query count by about 50%, so it's huge improvement). Results on Zen Cart Demo Store:

• without Query Cache: 950+ queries
• with Query Cache v1.5: about 200 queries

Download updated version here. Read my updated blog to find out more.

[evan70]

nice JOB work great at test site http://shop.kafeshop.cz ThnkZ

[yellow1912]

I wonder if this is the best approach to caching it. Why not editing the db class only, and then use the whole query string, md5 it to get a (fairly)u?

[data_digger]

I wonder if this is the best approach to caching it. Why not editing the db class only, and then use the whole query string, md5 it to get a (fairly)u?

I assume You're referring to changes in functions_categories.php?
Sure, it would be much more elegant way of doing this, but unfortunately (I tested it), it would be slow(parsing queries and argument extraction takes more time then executing it). In fact, I wanted to do it in similar way You wrote, but in performance contribution performance is more important then beautiful code.

[yellow1912]

I dont know why my comment got cut half way, anyhow what I meant is: if we treat each query string as an identifier (and by md5 the query content we would have a rather unique identifier for each query), then we can store each query result in a unique file?
I think the most time consuming would serialization and un-serialization of the object returned by the query (to save/restore the result), Maybe we can use some easier method to store it?

[data_digger]

I dont know why my comment got cut half way, anyhow what I meant is: if we treat each query string as an identifier (and by md5 the query content we would have a rather unique identifier for each query), then we can store each query result in a unique file?

Why would You like to treat query string with md5? I use whole query string as identifier (in array) in my class and it's quite fast.

Storing results in files is not a good idea imho. You never know if cached results are consistent with database. Also, 99% of queries execute in 0.01 sec or so (per query). Storing / restoring results of such queries would take much more time then executing it.
Besides, MySQL has it's own query cache (unfortunately hosting companies sometimes turn it off or forgot to turn it on) that does the same thing, only better, because results from cache are always consistent with database.