Forum Discussion
Help creating Channel opt in pivot / crosstab table
- 1 year ago
Hi klev28
I would recommend the "basket analysis" pattern for this.
See here for an explanation:
https://www.daxpatterns.com/basket-analysis/
For this pattern to work, you should unpivot the columns "is Email", "is Mail", "is Phone" so that you end up with this structure:
Customer Channel C001 Email C002 Email C002 Mail C003 Mail C003 Phone C004 Phone ... ... Then, if you follow the article, you should create two copies of a Channel dimension.
The # Orders Both measure from the article is similar to the one you're displaying in your example. To produce a matrix like your example, you would place one copy of Channel on rows and the other copy on columns.
Please post back if you need more help on the details.
Regards
- Anonymous1 year ago
Hi klev28 ,
I've made a test for your reference:
1\I assume there is a table(Table)
2\Add a caculate table
Channels = DATATABLE ( "Channel", STRING, { { "Email"}, { "Mail"}, { "Phone"}, { "Total" } } )3\Add 4 measures for the caculate table 'Channels'
Email = Switch( SELECTEDVALUE(Channels[Channel]), "Email",COUNTROWS(FILTER('Table','Table'[is Email]=1)), "Mail",COUNTROWS(FILTER('Table','Table'[is Email]=1 && 'Table'[is Mail]=1 )), "Phone",COUNTROWS(FILTER('Table','Table'[is Email]=1 && 'Table'[is Phone]=1 )), "Total",COUNTROWS(FILTER('Table','Table'[is Email]=1)) )Mail = Switch( SELECTEDVALUE(Channels[Channel]), "Email",COUNTROWS(FILTER('Table','Table'[is Email]=1 && 'Table'[is Mail]=1)), "Mail",COUNTROWS(FILTER('Table','Table'[is Mail]=1 )), "Phone",COUNTROWS(FILTER('Table','Table'[is Mail]=1 && 'Table'[is Phone]=1 )), "Total",COUNTROWS(FILTER('Table','Table'[is Mail]=1)) )Phone = Switch( SELECTEDVALUE(Channels[Channel]), "Email",COUNTROWS(FILTER('Table','Table'[is Email]=1 && 'Table'[is Phone]=1)), "Mail",COUNTROWS(FILTER('Table','Table'[is Mail]=1 && 'Table'[is Phone]=1)), "Phone",COUNTROWS(FILTER('Table','Table'[is Phone]=1 )), "Total",COUNTROWS(FILTER('Table','Table'[is Phone]=1)) )Total = Switch( SELECTEDVALUE(Channels[Channel]), "Email",COUNTROWS(FILTER('Table','Table'[is Email]=1 )), "Mail",COUNTROWS(FILTER('Table','Table'[is Mail]=1)), "Phone",COUNTROWS(FILTER('Table','Table'[is Phone]=1 )), "Total",COUNTROWS('Table') )4\Add a table visual
Best Regards,
Bof
Hi klev28
I would recommend the "basket analysis" pattern for this.
See here for an explanation:
https://www.daxpatterns.com/basket-analysis/
For this pattern to work, you should unpivot the columns "is Email", "is Mail", "is Phone" so that you end up with this structure:
| Customer | Channel |
| C001 | |
| C002 | |
| C002 | |
| C003 | |
| C003 | Phone |
| C004 | Phone |
| ... | ... |
Then, if you follow the article, you should create two copies of a Channel dimension.
The # Orders Both measure from the article is similar to the one you're displaying in your example. To produce a matrix like your example, you would place one copy of Channel on rows and the other copy on columns.
Please post back if you need more help on the details.
Regards
Hi Owen
Thanks for this. I've used the other solution for my Channel crosstab but the basket analysis that you highlighted is also very relevant and I will almost certainly be using this for the product page. It will save having to create a page per product group - which was also on my list of solutions to find.
I'm seriously delighted with both solutions 😄
Kezl