Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hi All,
I have an application have 350 million rows.I have 32 metrics in this app .
The user has an option to select an metric.
To do this, we have a inline load of all these 32 metrics.
On every selection of this from dropdown, it takes a lot of time.
Any suggestions on how to optimize this from performance standpoint?
Regards
Khushboo
May be you can use variable to get input from users using Input Box as an alternative?
Am I right that the values from the inline-table are only used to determine which dimensions/measure are displayed and not used within any of the calculations? Does similar charts but with the same but manually fixed dimensions/expressions react significantly faster than those one which are controlled by selections over the island table?
- Marcus
Hi Marus,
Not like that. When this values from inline load are selected to display values on Map, this action is caused by a button. When we click on button, it is extremely slow but again , but that is still OK.
I am just worried why that data island is slow? How to optimize that?
A data island is only useful for certain kinds of usability, for example to control the visibility of dimensions and/or measure or maybe some kind of filters within expressions, for example:
sum({< Date = p(DateValuesFromIslandTable)>} Value)
As soon as you used the fields from the island table within a "normal" expression like:
sum(Value * ValuesFromIslandTable)
Qlik will create a big temporary table with a carthesic product of all these fields and this needs a lot of resources and beside this it should be avoided like circular loops because the results are quite unpredictable and often not correct. This couldn't be optimized it needs to be replaced with a more suitable datamodel.
- Marcus
When it is getting evaluated in expression it is slow, understood.
When its just getting selected from drop-down, why is it slow?
Please take a look on the second question within my first answer - are similar objects really significantly more performant? Maybe there is a general performance issue and it's not directly related to the data island.
Further some more informations about where is which dimesion/expression in use and some screenshots from the object and datamodel would be helpful.
- Marcus