Comparing between different filters on the same data

[olik]

Say I have data like this:

Date	Account	Statistic	Value
30/06/2017	1	Alpha	0.85
30/06/2017	2	Alpha	0.34
30/06/2017	1	Beta	0.52
30/06/2017	2	Beta	0.29
30/06/2017	1	Gamma	0.14
30/06/2017	2	Gamma	0.25
31/07/2017	1	Alpha	0.35
31/07/2017	2	Alpha	0.23
31/07/2017	1	Beta	0.51
31/07/2017	2	Beta	0.26
31/07/2017	1	Gamma	0.16
31/07/2017	2	Gamma	0.90
31/08/2017	1	Alpha	0.88
31/08/2017	2	Alpha	0.20
31/08/2017	1	Beta	0.76
31/08/2017	2	Beta	0.79
31/08/2017	1	Gamma	0.83
31/08/2017	2	Gamma	0.66

Now imagine I have two bar charts (one per Account), with Date on X and Value on Y, and the different Statistics as different bars.

[In reality there are a lot more Accounts, and there are two slicers to select the two Accounts we want to compare.]

How can I get a third chart that shows the Value difference between the two, again per Date and Statistic?

I started by creating a duplicate of the dataset and now want to do something like https://community.powerbi.com/t5/Desktop/How-to-calculate-difference-after-different-filter-settings... - but there is no sensible way to relate the table with its duplicate. 

There must a way to do this surely?

[Thejeswar · Accepted solution]

Hi @olik,

You can try the below method as well!!! FYI, I have the data that you have provided for this purpose

The Following are the steps.

1. Have 2 copies of your data (namely Table1 and Table2)

2. Join the tables based on Date Column

3. Create the following 2 measures

In Table1

Measure 1 = CALCULATE(SUM(Table1[Value]), FILTER(Table1, Table1[Account] = SELECTEDVALUE(Table1[Account]))) 

In Table2

Measure 2 = CALCULATE(SUM(Table2[Value]), FILTER(Table2, Table2[Account] = SELECTEDVALUE(Table2[Account]))) 

Create another measure to find the difference between the measures

Measure = [Measure 1] - [Measure 2] 

Relationship, more likely to look like (not necessarily have to be same)

Relationship between tables through a bridge table

My Output looked as shown below

Output

 As you change the accounts in your slicer, you value changes correspondingly....

 Hope this helps you get what you needed!!!

regards,

Thejeswar

[v-jiascu-msft]

Hi @olik,

Could you please mark the proper answers as solutions?

Best Regards,

Dale

[Thejeswar · Accepted solution]

Hi @olik,

You can try the below method as well!!! FYI, I have the data that you have provided for this purpose

The Following are the steps.

1. Have 2 copies of your data (namely Table1 and Table2)

2. Join the tables based on Date Column

3. Create the following 2 measures

In Table1

Measure 1 = CALCULATE(SUM(Table1[Value]), FILTER(Table1, Table1[Account] = SELECTEDVALUE(Table1[Account]))) 

In Table2

Measure 2 = CALCULATE(SUM(Table2[Value]), FILTER(Table2, Table2[Account] = SELECTEDVALUE(Table2[Account]))) 

Create another measure to find the difference between the measures

Measure = [Measure 1] - [Measure 2] 

Relationship, more likely to look like (not necessarily have to be same)

Relationship between tables through a bridge table

My Output looked as shown below

Output

 As you change the accounts in your slicer, you value changes correspondingly....

 Hope this helps you get what you needed!!!

regards,

Thejeswar

[tiffanyt123]

Hello, I have 2 tables (coming from the same data source, I duplicated the tables and filtered by "version date"). I want to compare the data between them. Is it possible to use a DAX Formula to do so?

[olik · Accepted solution]

Thanks a lot! I used your suggestion and made the formula even easier:

Result = SUM(Table1[Value]) - CALCULATE(SUM(Table1[Value]), FILTER(ALL(Table1[Account]),Table1[Account] = SELECTEDVALUE(Table2[Account])))

That way I only need one measure, and technically I don't even need the full duplicate table... Just another list of accounts would do!

Do you see any issues with my approach? Still very new to the whole DAX coding!

[Thejeswar]

Hi @olik,

Your re-did DAX looks fine.

I think the new DAX that you have written should be good enough to get the difference

Regards,

Thejeswar

[olik]

I think I have found a way, using an additional measure based on SELECTEDVALUE parameters and then defined as such:

CALCULATE
(
AVERAGE(Table1[Value]),
FILTER(
ALL(Table1[Account]),
[Account]=Table2[Selected Account]
)
)

Surely that can't be the best way though?