Calculate

[scavalcantes]

I have the image table below, where I need to generate a custom column in the query editor with statistics (average, minimum, maximum, mode.sngl and etc) of the columns Gra1 through Gra5 and Esp1 through Esp5, but the calculation needs to disregard the zeros as we get using the excel formula averageifs.

[stretcharm]

You could either unpivot these columns and an calculate based on the attribute (Column) group.

http://radacad.com/pivot-and-unpivot-with-power-bi

or you could transpost pass the data to r which is very good at working with data sets like this and applying stats.

https://docs.microsoft.com/en-us/power-bi/desktop-r-in-query-editor

[scavalcantes]

The first option was to use the pivot, but the performace on a machine (i5, 6GB raw, 64bit and 1TB of hd, was compromised and when using dax union function to update only table 2018, simply the machine hangs.
Regarding the R still studying the possibility, but needs to be faster.
Would you have another solution?

[stretcharm]

You can also write multiple dax measures that calculate the stats for each of the base columns

https://www.daxpatterns.com/statistical-patterns/

I like to create a dax summary table than can be joined to your other data using ADDCOLUMNS and SUMMARIZE

https://www.sqlbi.com/articles/best-practices-using-summarize-and-addcolumns/

 BTW r is very fast, though does need the data to be in memory.

[scavalcantes]

O problema da performace foi devido a converter colunas em linha por grupo de colunas, com isso, o desempenho ficava compromentido.

A solução foi selecionar todas as colunas que tem valores a calcular e assim fazer o pivot