Hi, I have just made my first Direct Discovery load.
I have done a test case where I load approximate 10 millon rows. 3 dimensionfield and 2 Implicit/Measure field.
1 load Direct Discovery, 1 load the traditional way from a sql-server with an OLEDB connection.
The Loadtime is almost the same, nothing to care about. But when I select in a listbox and have a pivot with my 3 dimenssion and 1 expression it takes VERY long time to calculate the expression in the Direct Discovery application, not userfriendly. This is not even an option to use even if it saves a lot of space on disc. I have use 11.20 SR4 with the "old" Implicit statment. Have you experiensed the same as I have??
I have now test Direct discovery in 11.20 SR5. But One thing bother me, if i click in one of my dimension field and get the result from my Measure field, should not the value change the next time I click in the dimension field. This is value that changes in realtime for almost every second. Now I have to reload or close/reopen qlikview to get a new result and thats opposite the hole idea with direct discovery or I´m I missing something here?
Direct Discovery uses also a cache for the results obtained from the Direct datasource. So if you run the same query before the cache expiration time, QlikView uses the cache instead of running again the query against the data source.
In case you want decrease the cache time, try modifying the system variable DirectCacheSeconds in the script (more detailed info in the QlikView help). By default it is set to 1800 seconds (30 minutes). I think if you set it to 0 seconds, you will avoid the cache, and every time you click QlikView will run a query.
aha, thats why it works "sometime" probably after 30 min when I have done something else. I have read the FAQ and not notis about the cachesettings. I will read it again and do a new test with all the settings I need.