On Tue, Apr 09, 2002 at 08:06:08PM -0700, Zachary Buckholz wrote: > I am having a hard time with a logic problem related to scheduling an event > based on data stored in a MySQL table. Below is a sample data set. > The basic concept is a monitoring program that checks a website to see if > its up or down based on info stored in a table. > The way I am doing it requires me to know the last time the site was checked > and how often the site should be checked. I threw in the nc_epoch (next > check epoch seconds) while trying to work through this. That field can be > ignored, unless you see a solution using it. > > mc_id = customer id related to another table > url_id = url_id related to another table > chk_freq = how often in seconds the site should be checked > url_timeout = how long to wait for website before considering it down > lc_epoch = last time the site was checked - converted to UNIX epoch in DBI > query > active = is the site supposed to be checked > nc_epoch = next time the site should be checked - converted to UNIX epoch in > DBI query
If you have a column that says when the next check should occur, that's probably the column you should use: select * from table where nc_epoch <= now And then, for each of the rows: update table set lc_epoch = now, nc_epoch = now + chk_freq where mc_id = ? and url_id = ? Of course, you can ignore nc_epoch and just use lc_epoch and chk_freq if you prefer: select * from table where lc_epoch + chk_freq <= now update table set lc_epoch = new where mc_id = ? and url_id = ? I don't know if you can index the sum of two columns, however. Ronald
