Basically, I want:
1. Good concurrency / decent performance
2. Data integrity
3. Fast Search
4. Ability to backup per user
Backing up data by user is required for my solution. A lot of times, users screw up and they want to rollback to a previous state.
If I were to do a database per user, the backup/restore would be very straight-forward. I believe backup/restore procedure is similar for schemas (let me know if I am wrong here)? If I were to do a single schema/database, is it possible to get data per user and back it up? Select user rows, copy to a temp table/db, backup?
Thanks,
Frank
On Sun, Nov 15, 2009 at 1:11 PM, John R Pierce <pierce@xxxxxxxxxxxx> wrote:
undisclosed user wrote:depending on how many tables, etc, I suppose you could use a seperate series of SELECT statements ...
If I were to switch to a single DB/single schema format shared among all users , how can I backup each user individually?
but if this is a requirement, it certainly puts constraints on how you organize your data. without a much deeper knowlege of your application, data, and requirements, its kind of hard to give any sort of recommendations. you mentioned myISAM, so I gather this data isn't at all transactional, nor is relational integrity a priority.