You have to write the query so it gives the appearance of a crosstab.
So instead of:
select month,area,sum(sales)
from table
group by month,area
order month
you’d write
select month,
sum(case when area = 'A' then sales end) as sales_a
sum(case when area = 'B' then sales end) as sales_b
sum(case when area = 'C' then sales end) as sales_c
group by month
order month
We have a limitation with QaaWS and we cannot create cross tabs with the same. You might look at adding formulas and rearranging the data after its pushed to Excel.
Through QAWS ,bring all the fields and then build the logic in xcelsius to create a cross tab by using Match() and Index function or you can use vlookup().
If the number of rows pulling are less then, i really suggest to go with the excel functions in xcelsius or else you could go with the Mak1 by creating the derived tables in the universe level and then build the QAWS
Do as much as you can with your data before it gets anywhere near xcelsius. You can pull in huge chunks of data and manipulate it with vlookups etc, but you really are storing up problems for yourself in terms of performance and maintenance.
As I suggested and Mak 1 elaborated, create objects in your universe to build case statements for each total you require, so that the generated query SQL looks similar to my example. Basically you need to generate SQL which will always return a fixed number of rows and columns - this is the essence of a crosstab.
QAAWS is very limited when compared with BIWS. Why not create your Cross Tab in WEBI and then publish the block as a BI web service. Use the QAAWS connection in Xcelisus.
Hi Mark,
Java version ‘jre1.6.0_03’
but, how I know server side Java version and one more question why we need java version here… we are using rich client right… so ?