Question

Native Derived Table Count Values are Incorrect

  • 2 November 2017
  • 0 replies
  • 96 views

I have a user view and a login view - the login view contains a ‘count’ measure. I have created a user explore which joins in the login view:



view: user {

dimension: user_id {}

dimension: email_address {}

}



view: login {

dimension: user_id {}

dimension: date {

type: date_time

}

measure: count {

type: count

}

}



explore: users {

join: login {

type: left_outer

sql_on: ${user.user_id} = ${login.user_id}

relationship: one_to_many

}

}



I have one particular user who has logged in eight times. I have filtered the user explore for the user_id to verify this:



user_id | Login Count

06745456|8



This count measure is used in a native derived table (the labs feature introduced in 4.18):



view: user_metrics {

derived_table: {

explore_source: users {

column: login_count { field: login.count }

column: user_id {}

}

}

dimension: login_count {

type: number

}

dimension: user_id {}

measure: count {

type: count

}

}



I would expect the user_metrics explore to return the same results here as it does in the user explore but it does not:



user_id | Login Count

06745456|4

06745456|2

06745456|1



Why does it return three results instead of one? Additionally, why don’t those numbers add up to 8? I’m doing this in order to view a graph where the x axis is the number of logins and the y axis is the number people who have logged in that many times. For example, if I have one user who has never logged in, two users who have logged in once, and one user who has logged in twice:



4|

3|

2| *

1| * *

_________

0 1 2



Why doesn’t the query in my user explore match the user_metrics explore? The user_metrics explore should return the results that the user explore does.


0 replies

Be the first to reply!

Reply