tony <tony_at_email.com> wrote in message news:3C1A41B5.2070104_at_email.com...
>
> I'm thinking to make RDBMS itself with C or C++ language.
> Before starting, I already got problems...
> SELECT * FROM table1;
> As you guys know, the statement above retrieve all records in the
> table1. It's easy to retrieve from memory, but if the table1 has
> 1,000,000 rows
> and 10 columns, what RDBMS should do with that huge table ?
> The RDBMS should put all records in the memory ? 1,000,000 rows & 10
> columns r fine, but if the table has 1,000,000,000 & 1,000 columns ?
> This is totally wasting memory space if the RDBMS put the all records
> in to memory.
>
> anybody knows how RDBMS treats those tables(files) ?
The rule of thumb is that everything is stored on disk unless it is an
intermediate result and you are sure it fits comfortably in your memory.
Received on Mon Dec 17 2001 - 03:31:24 CST
