I read in an article that GET SNAPSHOT needs instance attachment whereas any SNAPSHOT TABLE FUNCTION needs database connection.What is the reason behind this difference??

Because the Table function definitions are stored in the database and SQL statement has to be executed to get the snapshot.

Cheers
Sathyaram

So,the tables functions are stored as metadata ?? What about the info for the GET SNAPSHOT FUNCTION..They are also stored in tables right??

http://publib.boulder.ibm.com/infoce...c/r0001449.htm

This is the API that gets snapshot information.

The GET SNAPSHOT command and the TABLE FUNCTIONS (which is EXTERNAL LANGUAGE C, BTW) are two different interfaces that call this API to get the information.

All table functions(whether snapshot or not) has to reside in a database and can be called using SQL. In the case of snapshot functions, the table function is an EXTERNAL LANGUAGE C function. The C program calls the snapshot API. Hence, to get the snapshot you have to connect to the database.

In case of GET SNAPSHOT command, the program for calling the snapshot API is invoked directly(ie, without going through SQL) and hence does not need a database connection.

In both cases, the instance snapshot information is held by DB2 in the same place (NOT inside the database).

These are just two different methods for accessing that information.


HTH

Sathyaram

Thanks for the information.Where does the information(counters) of the snapshot monitors get stored ??Does it get stored in system monitor heap??