|
Victor posted:There's also the issue of needing to verify data integrity/quality before you shove it in a production table. Very, very true. Keep an eye out for fields being cropped and row separators not performing correctly. I use the BCP tool a fair amount in ASE and SQL2005, I would recommend always using a format file rather than the -c switch. It's slight more laborious but it's helpful when you are trying to import a CSV (or what have you) that was supplied rather than generated personally. Also, when working in Sybase IQ, the LoadTable command is a total bitch, try and avoid where at all possible.
|
# ¿ Dec 20, 2007 12:30 |
|
|
# ¿ May 7, 2024 05:03 |
|
dangerous.hotdog posted:I'm having a hell of a time with a particular query. There is a USERNAME and SYSTEMDATE field and I need to develop a query that filters out records with a USERNAME/SYSTEMDATE pair that falls within 30 seconds of another USERNAME/SYSTEMDATE pair (with identical USERNAME). Basically, if a user connects again within the same 30 second window of their first connect, we want to omit that record in our analysis. Assuming i've not missed something very obvious... Yes. If you are running on a fairly decent box it would take 5-10 minutes tops with a couple of procedures/functions nested. What if they try again after thirty seconds, you want that kept? Here is how I would do it... Assuming can create a temp table. code:
|
# ¿ Sep 18, 2008 14:30 |
|
Triple Tech posted:Or do you keep the goons table generic and make tables to join off to like goons_tfr, goons_gws, etc that consist only of the goon pk plus the exclusive attributes? Without a shadow of doubt, this is the way to do it.
|
# ¿ Sep 18, 2008 18:19 |