hisouka:data
more accurately, for database, there is really no straightforward asnwer to this. if your db structure is currently of top quality and highly optimized, I dont see some ways for you to be able to reduce the size of your data. however, if there is room for improvement, I can think of several ways:
- use of proper data types. for example,
- use VAR* when necessary. Using VARCHAR(500) vs CHAR(500) will probably allow you to save on space. But of course there is minimal trade off against fragmentation and probably performance.
- use TINYINT instead of INT if you are sure that your data will be in the range of TINYINT
- NON-UNICODE vs UNICODE. Do take note that each unicode character takes up 2 bytes.
- use of bitmapping
- normalize when necessary to avoid several instances of the same information in your database
hisouka:trans log
I typically set a maintenance routine for my databases to have its transaction log backup every 30 minutes and have the database AUTO SHRINK property set to ON. This will continously shred off xlog and it wont grow that much. Use this settings with precaution as they might impact your DB's performance.
If you dont benefit from xlogs, you can turn its Recovery Model to Simple.
Or you can simple use the SHRINK command.
Hope this helps!