named set questions

As I understand it, a named set is not processed until it is needed.

Once these are processed are they cached for use by other client requests?

If there are Aggregations set up in the cube will this trigger the named sets be processed right away with the cube?

I assume that if a named set is too big there could be loss in performance, is there a way to hold down the size of a named set?

here are some examples of named sets i've created using some calculated members. perhaps there is a better way of doing this?

*************** Calculated Members***************************

CREATE MEMBER CURRENTCUBE.[MEASURES].[HP Plus Addl Billing]

AS aggregate([PREP CONTROL HDR].[Bill Formats].&[19],[Measures].[Billed Sales Amount]),

FORMAT_STRING = "Currency",

VISIBLE = 1;

2nd calculated member

CREATE MEMBER CURRENTCUBE.[MEASURES].[Prep Billing]

AS AGGREGATE(EXCEPT([PREP CONTROL HDR].[Bill Formats].[Bill Formats] ,{[PREP CONTROL HDR].[Bill Formats].&[19],[PREP CONTROL HDR].[Bill Formats].&[7]}),[Measures].[Billed Sales Amount]),

FORMAT_STRING = "Currency",

VISIBLE = 1;

*******************named sets************************************

CREATE SET CURRENTCUBE.[CTP Customers]

AS FILTER([CUSTOMER JOB].[Title Name].[Title Name].members,([Job Complete Date].[Calendar Full],[Measures].[HP Plus Addl Billing])> 0 );

CREATE SET CURRENTCUBE.[Prep Customers]

AS FILTER([CUSTOMER JOB].[Title Name].[Title Name].members,([Job Complete Date].[Year].&[2005],[Measures].[Prep Billing]) > 5000 ) ;

This one, I've used other named sets to create another named set.

Is this a bad thing?

CREATE SET CURRENTCUBE.[Non-Prep-Photo Customers]

AS EXCEPT([CTP Customers],{[Prep Customers],[Photo Customers]});

As I understand it, a named set is not processed until it is needed.

Once these are processed are they cached for use by other client requests?

Actually no - named sets are always processed and evaluated during MDX Script execution. This is done once, and they are cached afterwards.

|||

We've had issues where after the cube was processed it was not available for running reports until the server was rebooted.

When we removed the named sets the issue went away.

Is there any issues related to using named sets on a 32bit server running sql server 2005?