Improving the performance of a measure

[Anonymous]

Hello dear community first of all happy new year.

I have a purchase table like that:

Date	Client	Transaction Type	Invoice	Sales
01/01/2022 10:05	A	Redemption	A22	-10
01/01/2022 10:00	A	Purchase	A01	10
02/01/2022 15:00	A	Purchase	A02	15
03/01/2022 11:00	A	Purchase	A03	7

And my goal is to calculate the sum of sales after the first transaction (diferent invoince). In this case the result for customer A is 22.

For that I created three measures:

First Redemption = 
var _maxdate = MAX(Date[Date] ) + 1
return
CALCULATE(
MIN(Sales[Date]),
FILTER(
    ALL( 'Sales'),
    Sales[Date] < _maxdate &&
    'Sales'[Transaction Type] = "Redemption" )
)

Invoice = 
var _date = [First Redemption]
var invoice = 
CALCULATE(
    MAX(Sales[Invoice] ),
    KEEPFILTERS( 'Transacciones'[Date] = _data ) ) 
return invoice



Sales after fisrt redemption =
var _date = [First Redemption]
var invoice = [Invoice]
VAR __Sum =
        CALCULATE(
            [Sales],
            FILTER(
                    all(Sales),
            Sales[Date] > _data &&
            Sales[Invoice] <> invoice),
            VALUES('Sales'[Client]) )
 
RETURN if(isblank(_date), blank(), __Sum)

This is getting the right result, but it really takes a long time.

Any suggestions?

Thank you very much,
Simao

[parry2k]

@Anonymous Cool. Glad you have a solution in place. Cheers!!

✨ Follow us on LinkedIn and  to our YouTube channel

Learn about conditional formatting at Microsoft Reactor

My latest blog post The Power of Using Calculation Groups with Inactive Relationships (Part 1) (perytus.com) I would ❤ Kudos if my solution helped. 👉 If you can spend time posting the question, you can also make efforts to give Kudos to whoever helped to solve your problem. It is a token of appreciation!

⚡ Visit us at https://perytus.com, your one-stop-shop for Power BI-related projects/training/consultancy.

[parry2k]

@Anonymous glad it worked but just wondering what you mean by ending up using measure? I thought you were doing measures until now, no?

✨ Follow us on LinkedIn and  to our YouTube channel

Learn about conditional formatting at Microsoft Reactor

My latest blog post The Power of Using Calculation Groups with Inactive Relationships (Part 1) (perytus.com) I would ❤ Kudos if my solution helped. 👉 If you can spend time posting the question, you can also make efforts to give Kudos to whoever helped to solve your problem. It is a token of appreciation!

⚡ Visit us at https://perytus.com, your one-stop-shop for Power BI-related projects/training/consultancy.

[Anonymous]

@parry2k  For me it was indifferent whether it was a measure or a calculated table, I just wanted the most efficient way. @v-kkf-msft proposed a calculated table and I turned it into a measure

[v-kkf-msft]

Hi @Anonymous ,

Please try this measure to see if it works.

Measure = 
VAR FirstRedemption = 
    CALCULATE ( 
        MIN ( Sales[Date] ), 
        FILTER ( 
            ALL ( Sales ),
            Sales[Client] = MAX(Sales[Client])
             && Sales[Transaction Type] = "Redemption" 
        )
    )
RETURN
    CALCULATE ( 
        SUM ( Sales[Sales] ),
        FILTER ( Sales, Sales[Date] > FirstRedemption )
    )

If the problem is still not resolved, please provide detailed error information or the expected result you expect. Let me know immediately, looking forward to your reply.
Best Regards,
Winniz
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Anonymous]

Thank you both, @parry2k  and @AlexisOlson ,

But unfortunately it doesn't work yet. From what I've been researching the most efficient solution is to use ADDCOLUMNS - SUMMARIZE - FILTER, but I'm currently unable to pass the different first dates of purchase into the filter. The solution might be a measure or a table. It will look something like this:

(Table) 

Sales after First Purchase = 
VAR __firstRedemptionDate = [First Purchase] + TIME(1,0,0)
VAR _table =
     addcolumns(
         CALCULATETABLE(
         SUMMARIZE(
            FILTER(
                'Sales', 
                'Sales'[Date] > date(2021,11,1)),
        'Sales'[Client] ),
        filter(
            ALLEXCEPT('Sales', Sales[Date]),
            Sales[Date] > __firstRedemptionDate)),
        "Sales", [Sales]) 
return _table

But this way you also consider the first purchase.
Any other suggestion?

Once more thank you very much,

Simao

[AlexisOlson · Accepted solution]

If your [First Purchase] measure is reasonable, I'd expect a variation of this to work OK but I don't have anything to test it against:

ADDCOLUMNS (
    ADDCOLUMNS (
        SUMMARIZE (
            FILTER ( 'Sales', 'Sales'[Date] > DATE ( 2021, 11, 1 ) ),
            'Sales'[Client]
        ),
        "__firstRedemptionDate ", [First Purchase] + TIME ( 1, 0, 0 )
    ),
    "Sales", CALCULATE ( [Sales], Sales[Date] > EARLIER ( [__firstRedemptionDate] ) )
)

[Anonymous]

Thank you very much!

I ended up adding a SUMX and turning it into a measure, but it worked fine!

Thank you also @parry2k  for the help

[parry2k]

@Anonymous try this measure

Sales after First Redemption Date = 
VAR __firstRedemptionDate = 
    CALCULATE ( 
        MIN ( 'Table'[Date  ] ), 
        ALL ('Table' ), 
        VALUES ('Table'[Client  ] ),
        'Table'[Transaction Type  ] = "Redemption" 
    )
RETURN
 CALCULATE ( 
    SUM ( 'Table'[Sales  ] ),
    'Table'[Date  ] > __firstRedemptionDate,
    'Table'[Transaction Type  ] = "Purchase"
 )

You can tweak it as you see fit, make sure to change table and column name as per your model.

✨ Follow us on LinkedIn and  to our YouTube channel

Learn about conditional formatting at Microsoft Reactor

My latest blog post The Power of Using Calculation Groups with Inactive Relationships (Part 1) (perytus.com) I would ❤ Kudos if my solution helped. 👉 If you can spend time posting the question, you can also make efforts to give Kudos to whoever helped to solve your problem. It is a token of appreciation!

⚡ Visit us at https://perytus.com, your one-stop-shop for Power BI-related projects/training/consultancy.⚡

[AlexisOlson]

It appears that OP has a date table, so I'd encourage doing date filtering on that table instead. Having the time component in 'Table'[Date] makes this more complicated though.

@Anonymous, I'd encourage you to split your datetime column into a date column and a time column in your model for best performance. See this article for more detail:

https://radacad.com/how-to-use-time-and-date-dimensions-in-a-power-bi-model