C
ChUIMonster
Guest
You are reading twice as many records when you add "address" to the query so 2x makes sense. The fragmentation seems to be about 20% -- so that's another 20% -- still a far cry from 20x. I'll be surprised if dumping and loading has much of an impact. If "buffer hits" (or the lack thereof) is truly the problem then you would normally expect that running the same query twice in a row would show a big difference in response time (the 2nd run should be much faster because the data is in the buffers). Have you tested that? Is that your experience? If that isn't the case then it is more likely to be related to the transfer of data between the app server and the client or something about the code that surrounds the query rather than the query itself. (Or some unrevealed aspect of the query being run on the app server that differs from what you are telling us.) Your colleague is also querying roughly 5,000 x 2 records? How long does that take? Are we talking fractions of a second? A second or two? Minutes? Hours?
Continue reading...
Continue reading...