Ya, the front page of evolt is actually using blockfactor to speed it up, and it helps a *ton*. It's too bad it isn't documented better, cus a lot of people could probably speed up their queries by using it. It took a bit of playing around with, but I found that a blockfactor set to 75 worked the best for us, but thanks for the 32k reference, another hidden gem no doubt :) .djc. Raymond Camden wrote: > FYI, speaking of CF queries, are people here using BLOCKFACTOR any? It's > a _very_ easy way to speed up query time, and, unfortunately, is not > documented well at all. > > Basically (and anyone feel free to correct me on this), when CF performs > a query, it only fetches one row at a time. So, a query that returns 200 > rows will result in 200 fetches. CF can be told to get more at a time by > using BLOCKFACTOR. BLOCKFACTOR is a number between 1 and 100. The > default is 1. > > Ok, so here is where things get interesting. Why don't you just set > blockfactor to 100? Because CF, by default, can only get 32k of data at > one time. So, if 100 rows > 32k, CF has to 'back down' a bit in order to > fetch the right amount of data. What someone told me, was to simply take > the average row size and determine how many would fit into 32k.