Potential
performance improvements by splitting the data to be processed into groups and
simultaneously running multiple instance of your program to deal with
different groups of data.
If you have a one program that uses a temporary table and is
invoked multiple times, that single temporary table could be used concurrently
in multiple executions of the code. This could create unpredictable results
since the different instances of the code would be issuing delete, Inserts
and/or updates unsynchronized with each other.
You could solve the problem by creating multiple temporary
tables as a pool of tables. Each invocation of your program would have to
allocate an unused temporary table,
mark it as ‘in use’ , use it and release it back to the pool when you
through with it (for each Application Engine program you write).
