I can't see anything wrong with your fractile number (looking at the complete data set).
What do you expect to see?
I compared the outcome for example for the 2-quantile with the median(VENDA) as well as with
and got returned 44,4 (as with your table expression).
Also the other two values seemed reasonable to me.
I noticed that you are showing a cumulative relative (percentage) of the sum(VENDA), though I believe you shouldn't look at e.g. the 50% threshold VENDA value while looking for the 2-quantile / median, these are different metrics.
I agree with Stefan. I cannot find anything wrong with the fractile calculation.
I made a histogram to analyze the data, and it shows that 0, 44.4 and 165.7 are the fractiles you search for. I hard-coded these numbers in a calculated dimension and got this bar chart:
In it you can see that very few values are below 0, roughly 27% are exactly 0, etc. So the fractiles are correct.
Stefan and Henric,
Thank you for your answers and correct me on the usage of this function. My mistake .
I try to use this function to get the mininum value of sales on each quartile. I´ve created the cumulative % of Sales to check if the resulting value of the fractile column was correct.
Now I think it´s worth to use some sort of combination of rank , min and aggr functions to get what I need.
Pablo Labbe Ibaceta