Filter multiple rows against another table using less than

[GregCet]

Apologies for the awkward title, I am struggling to work out how to do this or even word the problem. I have a feeling it should be fairly straightforward but I am hitting a roadblock.

So I'm trying to write the basic form of a measure to look at the sum of a value before a product was launched. So the pseudocode for it would be:

SUM Number of Contacts WHERE ContactsData Year-Week < ProductLaunch Year-Week.

I added a column to both tables which has the year week in a whole number format of yyyyww (e.g 202123). Here is the DAX code for what I have so far:

Text

Pre-Launch Contacts = 
VAR ProductYW = FIRSTNONBLANK('Product Activity'[YW],0)

RETURN

CALCULATE(
    SUM('CS Data'[No of Contacts Answered]),
    FILTER('CS Data','CS Data'[YW] < ProductYW)

This currently doesn't work and just shows blank. HOWEVER, if I change the < to = then it does work, accurately showing the sum of contacts for the product activity year-week.

My thinking is that it has something to do with the fact that in the Conacts Data table there are lots of rows which will be having the same year-week but that is why I am using the filter method since it should be summing all contacts where the year week is less than the product year week.

Any ideas?

[MFelix]

Hi @GregCet ,

Don't know how you are presenting the information in your report but believe this is related with the type of filtering you are doing, that can be impacted by other context (filters, columns, values).

Try the following change to your formula:

Pre-Launch Contacts = 
VAR ProductYW = FIRSTNONBLANK('Product Activity'[YW],0)

RETURN

CALCULATE(
    SUM('CS Data'[No of Contacts Answered]),
    FILTER(ALL('CS Data''CS Data'[YW]),'CS Data'[YW] < ProductYW)

[GregCet]

Hi Miguel,

Thanks for your suggestion.

The syntax didn't seem to work when copying in your solution. I tried to change it to this but this produces the same results as before:

Pre-Launch Contacts = 
VAR ProductYW = FIRSTNONBLANK('Product Activity'[YW],0)

RETURN

CALCULATE(
    SUM('CS Data'[No of Contacts Answered]),
    FILTER(
        ALL('CS Data'),
        'CS Data'[YW] < ProductYW
    )
)

[MFelix]

Hi @GregCet 

Can you please share a mockup data or sample of your PBIX file. You can use a onedrive, google drive, we transfer or similar link to upload your files.

If the information is sensitive please share it trough private message.

[GregCet]

Thanks.

You can access two sample data files here. One for the contacts and one for the product activities:
https://drive.google.com/drive/folders/194g6y77zCW6Bs-4M2hidmSpzD_rn1HQW?usp=sharing

Please let me know if this does not work.

[v-yiruan-msft]

Hi @GregCet ，

I tried to create a table visual which apply your measure [Pre-Launch Contacts] and the field [YW] of table Product Activity on it base on your provided excel data, it can return the result just as shown in below screenshot.

How did you set the visual? Did you put the field [YW] of table Product Activity on your visual?

Best Regards

[GregCet]

Hello @v-yiruan-msft ,

Thanks for your support, I really appreciate it.

The visual I am trying to create is as follows:

• Each row should be the product activity description
• Columns for each row will show the PA launch week (from PA table) and then the sum of contacts from the contacts table from before the launch week.

As I said, I can get this to work when making a measure which looks at the sum of contacts that is EQUAL to the launch week but not a less than or greater than.

[MFelix]

Hi   @GregCet .

First of all looking at your data you have an issue with the YW column in the Product activity you have the week 8 and the value is 20218 and for the week 10 forward you will have with 202120 so has you can see the size of the number are different so you can have difficulties comparing for example products launched in the first 9 weeks of a year with the previous values.

For this you need to have the YW with 6 digits so it would be 202108.

I have a question regarding the date of the launch in this case your lower date is 23rd february in week 8, however your contacts table show information starting on week 17 how are you identifying the product launch?

[GregCet]

Hi Miguel,

Thanks for the response. I have amended the Year-Week columns in each table (of my actual PBIX file, not the sample) and this has not solved the issue unfortunately. To be clear, I padded the week number so that it would display as 6 digits (e.g. 202108 or 202006).

I did also realise after creating the sample that there was not much data which matches with the product launches. I only included a small sample since I am working with data on a weekly basis going back to the beginning of January 2019. I have now amended the sample file for product activities.

[MFelix]

Hi @GregCet ,

Still not understanding what is the launch date your product activity starts on week 17 and yuor first contact is also on the week 17 so no contacts before the first activity.

Sorry for making this questions but I'm not understanding the model to what you consider the week launch of the product that is the point that you need to compare to make all the counts that you need.

[GregCet]

I had made a bit of an error in the sample files provided as I limited some of the data which was available.

In my real world file, I have weekly data which goes back to January 2019 (201901) and so the measure should be summing all contacts made in the weeks from 201901 until the product launch date (e.g. 202117).

As I have said before, the week comparisons work when I use a = but if try to find a sum of contacts before (<) or after (>) then i do not get any results.

[MFelix · Accepted solution]

Hi @GregCet ,

I have copy paste some lines on the contacts and changed the year this gave me the folllowing values per week:

The cumulative values are the ones until that week (excluding current week) this table was made only making the use of the contacts table now making use of the measure I have sent earlier:

Pre-Launch Contacts = 
VAR ProductYW = FIRSTNONBLANK('Product Activity'[YW],0)

RETURN

CALCULATE(
    SUM('CS Data'[Contacts Answered]),
    FILTER(ALL('CS Data'[YW]),'CS Data'[YW] < ProductYW))

You can see that when you cross filter this with the YW of the product dates you get the results you need.

[GregCet]

Thank you for all of your time and effort Miguel! Turns out I had the same measure as you have in your example solution and it turns out that I had a stray relationship connection which was not necessary but was messing up the measure showing the correct result (for some reason).

Thank you again!