PHWinfo - Afficher un message - Doing Summation on multiple criterias on the same column in a singlequery

24/03/2008, 19h04

>> What if the fact IS a flag, and is not based on other data in your database? <<

It might or might not be based on other data; that is the problem. I
don't want to write triggers to constantly update the flags.

>> e.g. did the patient visit Congo anytime between 1978 and 1989? Instead of a "flag" saying yes or no, would you rather store each patient's entire travel itineraries for the period in question? <<

The alternative is a set of flags that cover all of the countries on
Earth, and all of the possible time periods

Oh, I also hope that
you created exactly the right flags, since they carry so little
data.

Eventually, I will probably have to have his travel itineraries to
follow a disease pattern. I have a friend who is an epidemiologist at
the CDC and that is how they do it.

>> What do you when that is not available, but the patient does know that he/she was there during that time period? <<

You put it the chart as free-form data and then you check his passport
records. Hey, this is the usual missing/fuzzy data problem. What did
you do when you never had a flag for that "bit" of data? What do you
do when the important year was 1990? More bit flags? Have you ever
considered what the triggers and constraints to keep the flags
current?

24/03/2008, 19h04	#6
--CELKO-- Messages: n/a Hébergeur:	Re: Doing Summation on multiple criterias on the same column in asingle query >> What if the fact IS a flag, and is not based on other data in your database? << It might or might not be based on other data; that is the problem. I don't want to write triggers to constantly update the flags. >> e.g. did the patient visit Congo anytime between 1978 and 1989? Instead of a "flag" saying yes or no, would you rather store each patient's entire travel itineraries for the period in question? << The alternative is a set of flags that cover all of the countries on Earth, and all of the possible time periods Oh, I also hope that you created exactly the right flags, since they carry so little data. Eventually, I will probably have to have his travel itineraries to follow a disease pattern. I have a friend who is an epidemiologist at the CDC and that is how they do it. >> What do you when that is not available, but the patient does know that he/she was there during that time period? << You put it the chart as free-form data and then you check his passport records. Hey, this is the usual missing/fuzzy data problem. What did you do when you never had a flag for that "bit" of data? What do you do when the important year was 1990? More bit flags? Have you ever considered what the triggers and constraints to keep the flags current?