all() doesn't work in filter formula

[Anonymous]

Hi, 

I created a small sales volumen table (fact table), this table is linked to order date table (dimension table). I then created a measure as below, which works normally.

sales vol Jan = calculate(sum(sales[sales amt]),filter('order date', 'order date'[order month]="2019-01"))

however, something strange happened when it add all() to the filter formula. 

sales vol January3 = calculate(sum(sales[sales amt]),filter(all('order date'), 'order date'[order month]="2019-01"))

It looks like the all() has caused a problem here, not sure about the reason. 

[Greg_Deckler]

@Anonymous Would need to understand your source data and expected output but yes, ALL can dramatically alter a calculation because you are essentially ignoring all filter context that have been placed on that calculation like column you are in, row you are in, any visual filters, etc.

[Anonymous]

Here is the sales table. 

Order DateShip Datecategorysales amt

2019-01-01	2019-01-01	coffee	3
2019-01-01	2019-02-01	tea	2
2019-01-01	2019-02-01	popcorn	3
2019-01-01	2019-02-01	chocolate	4
2019-02-01	2019-02-01	coffee	4
2019-02-01	2019-02-01	tea	4
2019-02-01	2019-02-01	popcorn	4
2019-02-01	2019-02-01	chocolate	4
2019-03-01	2019-03-01	coffee	4
2019-03-01	2019-03-01	tea	4
2019-03-01	2019-03-01	popcorn	4
2019-03-01	2019-03-01	chocolate	4
2019-04-01	2019-04-01	coffee	4
2019-04-01	2019-04-01	tea	4
2019-04-01	2019-04-01	popcorn	4
2019-04-01	2019-04-01	chocolate	4

It's linked to 'order date' table via 'order date' column. Below is a screen shot of 'order date' table. 

[Anonymous]

I found the reason.  I have created a new column called sales vol Juanary3. I meant to create a new measure, which is clearly not the case. I should have checked the fields panel more carefully. 

[CNENFRNL]

@Anonymous Even if you author a measure as such, the result still unreasonable.

As the FILTER(ALL(order date), ... ) returns a table with a column 'order date'[order month]="2019-01", it overrides all other filter context from the matrix, say [order month]="2019-02", "2019-03"..."2019-12", thus every row of matrix will be filled with sum(sales[sales amt]) of "2019-01".