Hi GD Team, I'm running into a bit of a problem conceptually and wondering if this is possible. I have a fact named Devices Seen that counts the number of mobile devices seen in a location. I want that value to be represented as a percentage, which is easy enough using the 'show in % button'. My issue now is I want to to show the Top 20 devices seen with an associated label (Home Local Government Area). Firstly I cannot use the show in %, and even if I did the %s would be calculated for the Top 20 and not all the Home Local Government Areas which would be incorrect. See the screenshot for an example. If the data was to be cut off at the top 20 mark I assume the percentages would be calculated as the top 20 being '100%' rather than all existing rows in the data. Is there a way I can implement a visualisation limit so that the data is still calculated over the full dataset for correct percentages but only the top 20 are shown?

Hi Paiwand, one option how to achieve this more complex condition is to use a MAQL metric. You can achieve the “show in percent” by dividing the standard metric by the same metric with “`BY ALL OTHER`” (meaning the total overall number). You can then achieve the filtering to TOP 20 with

HAVING TOP(20) IN (metric)

condition. The trick is to use

HAVING

instead of

WHERE

, this causes the filter to be applied after the inner metric is calculated and therefore it won’t affect the percentages and their base. So something like this should work:

SELECT
  SUM({fact/devices_seen}) / (SELECT SUM({fact/devices_seen}) BY ALL OTHER)
  HAVING TOP(20) IN (SELECT SUM({fact/devices_seen}))

Hi @Michal Hauzírek, first MAQL works perfect. Thank you!