D
Deleted member 147267
Guest
Yes, Ihsang, long data type allows over 2 billion values above zero so not likely you will run out of values. Replication ID consumes a LOT of space, and the only reason for using it IMO is if you want to guarantee that a value will be unique. However, that uniqueness was intended to be across any computer using a db that might create an ID and that value needed to be integrated into the main data set. This field will make your db several times larger than if you used an autonumber, and the difference will be exponential.
When I worried about users logged in, my code added them when the logged in, and removed the record when they logged out. The max of possible users was 25, so the max records I could have had in the table at one time would be 25. I didn't care when anyone logged in/out, only if they were in. If that's all you need, then you don't need tens of thousands of login records. Whenever they did anything crucial, I stored the userID as part of the creation/edit record. In short, to me it didn't matter when anyone logged in/out, just mattered if they were in and what they did when they were.
When I worried about users logged in, my code added them when the logged in, and removed the record when they logged out. The max of possible users was 25, so the max records I could have had in the table at one time would be 25. I didn't care when anyone logged in/out, only if they were in. If that's all you need, then you don't need tens of thousands of login records. Whenever they did anything crucial, I stored the userID as part of the creation/edit record. In short, to me it didn't matter when anyone logged in/out, just mattered if they were in and what they did when they were.