Problem: PL/SQL procedure that inserts into 1 of 4 tables

Consider the problem of storing data of variable type -- one table for integers, one for floats, one for strings, and one for dates. Each row in those tables has a key associated with it that refers to a table that stores the type of each datum:

META	key	value_name	stored_in_table
	1	a		core_integer
	2	b		core_string
	3	c		core_integer
	etc.

CORE_INTEGER	key	timestamp	value
		1	x/x/x		10
		1	y/y/y		20
		3	z/z/z		30

CORE_STRING	key	timestamp	value
		2	x/x/x		'foobar'

Now, I'd like to write a function that accepts two parameters, a value_name and the value itself, and stores that data in the appropriate table.

Right now, I built this as a if-then-elsif block that looks rather ugly:

select * into METAREC from META where value_name = param_value_name

if METAREC.storedin = 'core_integer' then

        insert into core_integer ...
elsif METAREC.storedin = 'core_float' then

        insert into core_float ...
etc.
you get the picture...

The problem is also that I should be able to create new META keys and store their data in arbitrary tables, which are only common in their field names, without having to change the procedure code.

Is there any way to dynamically create and run a PL/SQL block?

Also note that these tables may grow to the order of 300-500MB (yes, that's megabyte)...

Thanks,
trf

Union Bank of Switzerland       | Email: Frank.Tropschuh_at_ubs.ch
OE EDSZ/EZIU/EPPD TRF VC419     | Tel:   +41-1-236-9823
Performance Management          | Where: Flurstrasse 65, Geb. VF
Decentral Systems               |        CH-8048 Zurich


=================================================================