Hi all.
This seems similar as How to make histogram by grouped aggregation count? to me that is nicely answered by @timroes.
I have a data that I think has the same structure, although no integers here:
{ "action": "play", "user": "john", "content": "cnn" }
{ "action": "play", "user": "maria", "content": "cnn" }
{ "action": "play", "user": "john", "content": "bbc" }
{ "action": "play", "user": "maria", "content": "bbc" }
{ "action": "play", "user": "john", "content": "cnn" }
{ "action": "play", "user": "john", "content": "cnn" }
{ "action": "play", "user": "john", "content": "vh1" }
On that I create a table report to show me how many times the content is played using terms and split row
And this works nicely. For above data set this would give me:
cnn 4
bbc 2
vh1 1
But I also need another report to show the count of users that played this content. Since one user can play one content multiple times, I just want to see how many users have done the play action on content.
For above data set this would the look like:
cnn 2
bbc 2
vh1 1
I actually need just the report so I choose data table which I can export to csv. I tried to do some sibling aggregations but I always end up without grouping. Can this be done in some other way?
Thanks