Multiple datasets for each query or one dataset for all queries?

Bryan Leow
edited February 11, 2022 in Analytics #1
Suppose I have a chunk of queries that counts the number of times a transaction occurs in each warehouse and runs said query for all warehouses
elect transactionCount, warehouseID FROM warehouse2.itrn where...,
select transactionCount, warehouseID FROM warehouse3.itrn where...,
select transactionCount, warehouseID FROM warehouse4.itrn where...,
.
.
.
select transactionCount, warehouseID FROM warehouse117.itrn where...,select transactionCount, warehouseID FROM warehouse1.itrn where...,
s
(simplified version of query)

I then want to display the following data derived from this chunk of queries:
a) aggregate transactionCount from all warehouses
b) aggregate transactionCount grouped by warehouse country (e.g for all US warehouses, for all German warehouses)
c) top 10 warehouses with most transactions
Each of these a, b, c will be displayed on separate charts i.e 3 charts total


Would the query run faster if
1) I put the chunk of queries + queries for a), b) and c) all in ONE dataset and then link that dataset to all the charts, or
2) put chunk of queries + queries for a) in ONE dataset, chunk of queries + queries for b) in ANOTHER dataset ... etc

Or is there an even faster way than #1 and #2 in computing answers for a) b) and c)?

The most related thread I could find is https://forums.opentext.com/forums/support/discussion/42563/multiple-querys-vs-one-query-in-birt-report but it was from more than a decade ago (im currently on the latest birt 4.6.0 while the linked thread is on 2.3) and didn't go into detail