compsol1993
asked on
Automatic Sum in MySQL
Hello,
Please consider the following tables:
Table #1 Day:
ID - int key
Year - int
Month - int
Day - int
DayVal - float
Table #2 Hour:
ID - int key
Year - int
Month - int
Day - int
Hour - int
HourVal - float
Is there anyway to automatically sum all of the HourVals of Table #2 into the DayVal of Table#1 for a given day? I'd like this to be automatic. For example, if I poll DayVal at any time it always contains the sum of the values in the sub-table. I've been investigating using triggers but some sites have claimed this is not a good way to do this. Am I missing a standard method for doing this. Any direction would be appreciated. Just for perspective, this database is going to store approximately 1480 values for each hour.
To take this to the next level, I am also considering going down to data minute-by-minute, and having the hour value be a sum of minutes, but that is not as important right now.
Any direction you could offer on this would be appreciated.
Please consider the following tables:
Table #1 Day:
ID - int key
Year - int
Month - int
Day - int
DayVal - float
Table #2 Hour:
ID - int key
Year - int
Month - int
Day - int
Hour - int
HourVal - float
Is there anyway to automatically sum all of the HourVals of Table #2 into the DayVal of Table#1 for a given day? I'd like this to be automatic. For example, if I poll DayVal at any time it always contains the sum of the values in the sub-table. I've been investigating using triggers but some sites have claimed this is not a good way to do this. Am I missing a standard method for doing this. Any direction would be appreciated. Just for perspective, this database is going to store approximately 1480 values for each hour.
To take this to the next level, I am also considering going down to data minute-by-minute, and having the hour value be a sum of minutes, but that is not as important right now.
Any direction you could offer on this would be appreciated.
ASKER
Ok, it looks from these options that the view would be optimal. My only concern is the amount of data we are looking at. Right now the hourly data table contains about 14 million data entries, and that has only accumulated over a year's time. So next year at this time we can expect 14 million more.
I ran just one parameter's sum, and it took about 5 seconds. Is this normal, and is this still the optimal choice considering this amount of data. My thought was that a trigger would always have the data ready because it was calculated on INSERT, not on SELECT. For the end user, what you advise as the most efficient option.
I ran just one parameter's sum, and it took about 5 seconds. Is this normal, and is this still the optimal choice considering this amount of data. My thought was that a trigger would always have the data ready because it was calculated on INSERT, not on SELECT. For the end user, what you advise as the most efficient option.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
1) Use a trigger
2) Look for all occurrences in your code where you update HourVal and add more code to keep DayVal in sync
3) Have a periodic process keep generate the proper DayVals (sort of like an Oracle Materialized View). This will cause DayVals to be correct only as of the last time the process ran.
4) Get rid of Table #1 and use access it as a view of Table #2 (i.e. SELECT ID, Year, Month, Day, SUM(HourVal) FROM Table2 GROUP BY ID, Year, Month, Day