Counting codes across multiple fields and re-labeling the codes

I'm a new user and trying to do some code counting aggregation.

I have 5 columns of codes (Col1, Col2, Col3, Col4, Col5), and in each there are 20 possible code values for a data record, let's say the possible values fall within 00 to 20.  Only 1 possible code per record per column. So data might look like:
Rec  Col1  Col2  Col3  Col4  Col5
A       05     10     
B       12     00     19
C       01     12     15     00     05
D       05
I come from an excel and SQL background, but not sure how to crack this beast mode (or otherwise).
Aside from counting across all 5 columns so that the output indicates count of each distinct code (eg: three of 05, two of 12, two of 00, etc), the codes need to be relabeled into values.  So for simplicity, see attached, and imagine instead of the values 10, 09, 05, etc, it's something like Green, Red, and Blue.
Any direction would be appreciated, thank you.

Best Answer

  • AS
    AS 🔵
    Accepted Answer

    Welcome!

    What I'd do is bring your data into a Magic ETL dataflow and put it through the Collapse Columns transform to restructure it:

    Rec Value

    A      05

    A      10

    B      12

    B      00

    B      19

     

    Then in the card you can just simply count the occurences.  You could also join in a mapping chart for what each of these codes means in terms of color.

     

    Code  Color

    00       Black

    05       Red

Answers

  • You will want to collapse the columns for sure.  Does the column number mean anything as well?  Would it matter if col 1 was 05 over col 3 being 05?

     

    if so, you would want to collapse your data to have 3 fields;

    rec col value

    A    1     05

    A    2     10

     

    and so on

  • Thanks, this looks like it will do the trick!