Thread: Performance questions

Greetings All,

I have a couple of questions about performance.


1. I have an xmlpackage which queries the DB for the top 3 bestselling products in the past 7 days for a given category. I then display these on the category page. If there have been no sales (or less than 3) - then I'd like to extend the search to the previous 14days & then the previous 30 & then 60 etc - until I get 3 results. I realise I could achieve this by having multiple queries in the xmlpackage but I'm wondering what the performance implications are of this & if there's a better way?

2. I have an xmlpackage which displays just one product in a specific way - I call this xmlpackage from quite a few pages & just pass it the productid - what are the performance implications of this - let's say I call this same xmlpackage 12 times on the same page - I know this will result in 13 DB queries (the one for the page & then the 12 times this xmlpackage is called) - but is that a big deal?


I'd be grateful for your thoughts.


TTFN

BFG

Anyone have any thoughts?


TTFN

BFG

This was a typical homework assignment for studying how circuits are designed using Boolean algebra.
Once the Boolean diagram is made the programming is fairly easy and coherant.

I'm not sure I follow?
I'm comfortable with creating the code - I'm asking about the performance implications.............



TTFN

BFG

This is an intriguing topic, especially the math and logistics behind it (sorry I couldn't add anything of use).

We recently ran into the same question after building a fairly sophisticated template with lots of queries. What we ended up doing was calling a single stored procedure with parameters, and did the successive queries from there rather than multiple calls from the template.

Not sure if it is applicable, but that is where we did it.

I am a firm believer that the fewer calls to the database the better.

As the products probably only need updating daily, I would run an sp overnight and add the results to the category in extensiondata and simply read the products from there when you get the category info.

Obviously your method would be required if you need realtime info.

I think that if you did all of the number crunching on the database itself then you'd have very little if any performance hit. Especially if you're only running it once per day.