topic Fact table with several columns as measures in App Development

Fact table with several columns as measures

ali_hijazi — Thu, 27 Oct 2022 10:09:16 GMT

Hello
I'm reading 3 fact tables from a database
these tables have like 30 columns of which 10 are measures
these fact tables are linked via a link table in Qlik (not concatenated into one table)

there are measures that require summing diverse columns from the 3 fact tables; and this is causing low performance in the chart that display these measures

I was thinking of using cross table and minimize the measure columns into one; the thing is that the fact tables have so many rows (14 to 15 millions rows) so if I make a crosstable then I'm ending up with over 100 million records for each fact table
so now if I concatenate the resulting table, will there be degradation in performance?
kindly advise

Re: Fact table with several columns as measures

Mark_Little — Thu, 27 Oct 2022 10:25:04 GMT

HI @ali_hijazi

Performance depends a lot on the server spec and concurrent user base.

But i have worked on apps with a lot more data in that 14 or 15 million rows with no problems, as long as you have a good data model and a decent server you should be ok.

Re: Fact table with several columns as measures

ali_hijazi — Thu, 27 Oct 2022 11:18:49 GMT

The thing is that when I do a cross table on these tables, I end up with over 100 million records each

Then I need to concatenate these tables so I end up with a table with 300+ million records

And this number is subject to increase every month