Problem DAX measure

[Yikwun831]

Hi all, I am a beginer in Power BI, i have strugggling with a mearsure problem. For example,

Table 1 

Customer_ID	date	Register	Referral Link	First Visit	First Click into bag	First_Purchase	Online shop Channel
26	1/4/2011	0	0	0	0	1	7
26	11/7/2012	0	0	0	1	0	44
26	1/12/2013	0	0	0	0	1	17
26	14/3/2014	0	0	1	0	0	6

I want to calculate the first_Purchase(FP) for the Customer_ID by ignoring the online shop channel (obviously customer ID 26 his first_purhcase(FP) would be 1/4/2011 by ignoring the shop Channel

I only got the right answer once i created a new table by the following

step 1.

New Table = CALCULATETABLE(SUMMARIZE(table 1,table 1[Customer_ID], "Mdate",CALCULATE(MIN(Table 1[date]),Table 1[First_Purchase]=1)), Table 1[First_Purchase]=1)

step 2.

Nr of customer with at least 1 first purchase = count('New table'[Mdate])

The answer will be 1 (as i only want to get the min date of first purchase)

However, i used same logic to create the measure in table

e.g distinctcount(table 1[customer ID],first_purchase =1). or countrow(summarize(tabl 1..))

the result i get is larger than the New Table answer. it seem this in table 1 counting those record with different shop channels therefore this customer ID would be show again on Dec 2013 (What i want is just show once in april 2011).

I don't want to change the table structure it is because i have to do further analysis in all customer's first visit and click into bag etc..

Thanks and appreciate with your help

[Greg_Deckler]

Hi @Yikwun831, I am not sure that I follow all of this 100% but if you want a visual where you have the Customer_ID and min date of first purchase you should be able to create a measure like this:

Measure = 
VAR __customerID = MAX('Table 1'[Customer_ID])
VAR __table = FILTER(ALL('Table 1'),[Customer_ID] = __cucstomerID && [First_Purchase] = 1
VAR __date = MINX(__table,[date])
RETURN
__date

Put that measure into a visual with your Customer_ID and see if that is what you are looking to achieve.

[Yikwun831]

Thanks a lot Deckler, 

In this " __cucstomerID && [First_Purchase] = 1"

i don't have the selection of [first_purchase] (this column data i generated by query from database), so at the memont I still cannot test the measure yet.

In addition, I have loads of data for example customer ID 27,28 etc. Their each first action (first view, first purchase...) has been fully recorded and generated a row for each first action in different online shop channels.

So ultimately, i would like to show their first purchase by year month. 
for example, Customer ID 26 only show in april 2011

but not in both april 2011 and Dec 2013 (otherwise it will be duplicate my calculation for new customer because i am defining the totally new customer instead of a exist customer who having a new purchase in new channel but they have already purchase in another channels

Thank you so much Deckler!

[v-eachen-msft]

Hi @Yikwun831 ,

Here is my sample data.

Then I created a new measure to get date of first purchase.

Measure =
CALCULATE (
    MIN ( test[date] ),
    ALLEXCEPT ( test, test[ID] ),
    test[first purchase] = 1
)

Here is the result. Does this suit your needs?

[Yikwun831]

Yes, this measure is great! however, I want to count their first purchase instead of showing

 

For this example, i want is ID_26 will be only showed in April 2011, 27 only in May 2013 and 28 only May 2015.

It only happens in the month that has its inital first purchase (even they have another first purchase in other channel)

secondly, i want to visualize that measure no matter in matrix form / stacked chart.

I tried many ways and method, but finally it doesn't work.
the graph will show all the first purchase in different channles.

thanks a lot mate! it would be great if i have your help

[v-eachen-msft]

Hi @Yikwun831 ,

I am still a little ambiguous about your needs. Would you like to provide a sample result you want?
And you can download my test file, I created two kinds of matrix form in it.

[Yikwun831]

Thanks for your help. I am really appreciate that.

I have downloaded your sample and read through it, it was really good! it helps me to understand the DAX and logic behind this.

The matrix with measure 1 is the thing i want exactly (but the number of total of row&column seems all 0)

Furthermore, i would like to have the stacked column chart with x axis = year month (Date) and y-axis = sum of the no. of first purchase in that month (the "first purchase" has been defined in previously)

then using the same dax to create the new column like your table column 1 and 2, but it doesn't work in my tabke.

Hence,I am just wondering that might due to the edit queries problem? ( as I extract the data from database by SQL and then consolidate the data)

Basically, my first step is extracting data (the first view, first purchase etc all is timely data) by SQL, then i unpivoted columns and used conditional cloumn to create new columns for first purchase, etc.

Customer ID	Online shop Channel	Reg	first view	first click	first purchase
26	1	14/2/2011
26	2		14/5/2010
26	9				5/6/2011
26	8				18/12/2014
26	13			18/12/2010
27	3				9/9/2013
27	4			1/8/2013
28	5	1/1/2019
28	6		14/4/2018

then i unpivoted columns

Customer ID	Online shop Channel	Type	Date
26	1	Reg	14/2/2011
26	2	first view	14/5/2010
26	9	first purchase	5/6/2011
26	8	first purchase	18/12/2014
26	13	first click	18/12/2010
27	3	first purchase	9/9/2013
27	4	first click	1/8/2013
28	5	reg	1/1/2019
28	6	first view	14/4/2018

then conditional column

Customer ID	Date	Online shop Channel	Reg	first view	first click	first purchase
26	14/2/2011	1	1	0	0	0
26	14/5/2010	2	0	1	0	0
26	5/6/2011	9	0	0	0	1
26	18/12/2014	8	0	0	0	1
26	18/12/2010	13	0	0	1	0
27	9/9/2013	3	0	0	0	1
27	1/8/2013	4	0	0	1	0
28	1/1/2019	5	1	0	0	0
28	14/4/2018	6	0	1	0	0

after all, i started working with those data like creating the measure and visualize the date

is there any mistake happen so i cannot create the measure and column like yours? 

Thanks