Report on Column with Multiple Unique Values

Options
user00230
user00230 Member
in Archive

We are looking to build cards on a field/column that has multiple values from JIRA and need to report on it into DOMO.

 

Fi eld Information: Field contains associates names, and can contain anywhere from 0 to 20 people in this field for each row. This field can contain anyone from hundreds of end users we have today in the environment. 

  • For Example: Field contains values of : End User 1, End User 2, End User 3

 

When this field is pulled into Domo, each row containing this information in the column is now just one string of text and one value, instead of multiple values. We need to be able to build a card in DOMO to show the separate values in this field, and sum the total based on each end user: 

 

Example of Data Pulled Into DOMO: 

  • Row One, Field Value: End User 1
  • Row Two, Field Value: End User 1, End User 2
  • Row Three, Field Value: End User 1
  • Row Four, Field Value: End User 1, End User 3, End User 2

 

Example for Reporting Needed on Card: 

  • End User 1: 4 results total
  • End User 2: 2 results total 
  • End User 3: 1 result total 

 

Example of Reporting We are Getting Now (not what we need): 

  • End User 1 Total: 2
  • End User 1, End User 2 Total: 1
  • End User 1, End User 3, End User 2 Total: 1

 

What we want to be able to do is report how many times an end user shows up for each record, based on our process in the tool. How do we get DOMO to identify these unique values for the end users in order to report on each one individually? 

 

 

Tagged:

Comments

  • PodiumMason
    PodiumMason Contributor

    Hi there!

     

    I'm having a bit of trouble visualizing your dataset based on your row level description. Do you have column headers for each of the different values in the row? 

     

    For example: 

     

    Category           User 1              User 2

    Field Value        End User 1

    Field Value        End User 1      End User 2

     

     

    Could you possibly attach a screenshot of your dataset? That would make it easier to determine what transformations might need to take place to get the result you need. 

     

    Thanks!

     

     

     

     

     

     

     

     

     

     

     

     

     

     

     

    **Say 'Thanks' by clicking the thumbs up in the post that helped you.
    **Please mark the post that solves your problem as 'Accepted Solution'
  • user00230
    user00230 Member

    Thank you for the response.  The dataset would not have different column headers for each of the different values; it would all be the same column header. This field can capture up to 500 different end users, and we cannot feasibly have 500 different column hearers for each end user. 

     

    For example: 

     

     

    Record#       Users                                                        Date

    123               End User 1                                               5/31/2017

    124               End User 2, End User 1                           5/26/2017

    125               End User 3, End User 2, End User 1       6/01/2017

    126               End User 1                                               5/26/2017

  • DataMaven
    DataMaven Coach

    I wonder if you put all the columns - that are not that user column - to the left.  

    Then do a function to separate the values across columns - no matter how many that makes. 

    Then have the count execute on the array, rather than a single column.  

     

    I don't know, technically, how it would be accomplished in Domo, but I am hoping this may jog something for you that could help you keep going.  

     

    Good Luck!

    DataMaven
    Breaking Down Silos - Building Bridges
    **Say "Thanks" by clicking a reaction in the post that helped you.
    **Please mark the post that solves your problem by clicking on "Accept as Solution"
  • user03166
    user03166 Member

    Hi @user00230 ,

     

    Did you by any chance found the solution to this, I have a simmilar challenge.

     

    Thanks,

  • cwolman
    cwolman Contributor

    Assuming that there are 0-20 possibly values you could make a MySQL dataflow like the following:

     

    1st transform - named numbers

    select 0 as `no` union
    select 1 as `no` union
    select 2 as `no` union
    select 3 as `no` union
    select 4 as `no` union
    select 5 as `no` union
    select 6 as `no` union
    select 7 as `no` union
    select 8 as `no` union
    select 9 as `no` union
    select 10 as `no` union
    select 11 as `no` union
    select 12 as `no` union
    select 13 as `no` union
    select 14 as `no` union
    select 15 as `no` union
    select 16 as `no` union
    select 17 as `no` union
    select 18 as `no` union
    select 19 as `no` union
    select 20

     

    2nd transform - named data

    select `Record#`
    , trim(substring_index(substring_index(d.`Users`, ',', n.no), ',', -1)) as `User`
    , `Date`
    from `numbers` n
    join `dev_test` d on char_length(d.`Users`)-char_length(replace(d.`Users`, ',', ''))>=n.`no`-1

     

    ** Replace `dev_test` with the actual name of your dataset

     

    Output Dataset

    select `User`
    , count(`User`) as `value`
    from `data`
    where not `User`=''
    group by `User`

     

     


    -----------------
    Chris

Categories