Dax Calculate column even when there is no record at a date

Question

We are creating a powerpivot model based on a table with data about worktasks of our employees.
This is simplified our input:
Tasknr date started dateended
1 20140101 20140201 (yyyymmdd)
2 20140102 20140103
3 20140104 20140108
Etc
We created two measures to calculate how many tasks are opened and closed per day. That was no problem, but now the tricky part. We also want to know each day how many tasks are
still open, even on days that no tasks are opened (or closed).
My approach was to create a calculated column to determine how many tasks are opened up to that date and subtracted the closed tasks and end up with the tasks that are open. That
seems to be working :
=COUNTROWS(FILTER(Tasks; Tasks[Open]<=EARLIER(Tasks[Open]))) - COUNTROWS(FILTER(Tasks; Tasks[Close]<=EARLIER(Tasks[Open])))
When I load this to the excel pivot I end up with this:
Row Labels| Opened | Closed | Open not finished
1/1/2014
        1         
0          1
1/2/2014        
1          0         
2
1/3/2014
        0         
1         0
1/4/2014
        1         
0         2
The line with 1/3/2014 should have 1 in the column [opened not finished], but because that calculation is linked to the opendate column, not to the closed and 3/1/2014 no task
was opened, there is no value. On days that no tasks are opened or closed there even is no line at all although there could be tasks that are still open.
I need a mechanism that calculates the value even when there are no records
on a particular day. Our users want to be able to view the results on any date they select.
I have considered a second table with all the dates and calculate columns from the task-table but are there other ways to do this? I searched this forum but did not find an answer
so far.

Answer 1

Hi Jacob,
In our company we have a standard date table that is included in all of our models. What I didn't want to do in this case was to extend that table with calculated fields to solve this issue. Alternative was to create a new date table with the calculated
fields I needed. But I don't like that also so what I did was rewrite the sql that loaded the data into the pivot so that the measures were calculated at the load. But I am also not happy with that solution because of maintanance and performance reasons. My
feeling is that there must be a way to solve this with only dax in the loaded table.
Jacob's answer does exactly what you want. The DAX expression in his response is a measure which you could put in your Tasks table. You don't need to alter your date dimension in any way.
The key to this technique is that the date table cannot have an active relationship to your Tasks table. (although you could have an inactive relationship and then you could use the USERELATIONSHIP function to make other measures easier to calculate)
Translating Jocob's measure into something against a 'Tasks' table would looks like the following:
=
CALCULATE (
COUNTROWS ( Tasks ),
FILTER (
Tasks,
Tasks[Open] <= MAX ( Calendar[Date] )
&& Tasks[Closed] >= MAX ( Calendar[Date] )
http://darren.gosbell.com - please mark correct answers