General Domo Dojo Questions

Hi - I have a few quick questions!

 

Nulls: 

1. Would it be possible to assign the value 'null' to blank values on an imported data set? This would make it easier to work with those values when filtering, and wouldn't require us to map it to a string first. 

Tables:

1. Are there any plans to allow for hidden columns by column variable rather than just by column index? 
2.  It appears to be a glitch that when lables are created for column names and the column is moved or deleted, the label is re-assigned. When working with many columns, this can be very problematic and mess up the column labels and names significantly. Has Domo been looking into fixing this issue?

Best Answer

  • jaredpk
    Accepted Answer

    Thanks for the clarification.  For the nulls, if you are using workbench there is a search and replace option before you upload the data.  That's a possibility, but only if you are using WB to load the data.  If not, you might have better luck doing the ifnull logic in a dataflow or MagicETL rather than a beast mode.  You could then replace the nulls without having to create a beastmode column.

     

    For question 1 on the tables, that now is clear.  Unfortunately it runs on column numbers, but I think that's an excellent enhancement.  There is an ideas exchange here on the dojo to submit ideas or the feeback button in the product (checkerboard menu as if going to the data center or admin).

     

    For question 2 on tables, I have seen that as a problem if you are using appends and then change the structure.  With appends that can't be entirely avoided because there is an old and new structure co-exisiting.  Full replaces should not have that problem.  If there is that problem and it's not related to appending datasets then the best thing to do would be to show the specific problem table in a support ticket.

Answers

  • HI there -- regarding nulls, if I understand correctly you are asking for a string with the word 'null' to be automatically included.  There are some disadvantages to that.  When a value is null, that is literally the absence of information.  So calculations can run on cells with data, and will simply ignore the null cells.  If there was a string there, it would cause calculations to fail.  That said, there are some functions that allow you do do this in dataflows:

    MySQL / Beastmode -- use the ifnull function

    Redshift -- use the nvl function

    MagicETL -- use a replace values function and select null values in the magicETL.

     

    Regarding the tables questions, please provide a little more context on this.  Can you clarify question 1 in general to help me what you are trying to accomplish.

     

    Regarding question 2, are you using workbench or some other method to import the data in question?

  • Thanks for the swift reply! I've included my responses below:


    @jaredpk wrote:

    HI there -- regarding nulls, if I understand correctly you are asking for a string with the word 'null' to be automatically included.  There are some disadvantages to that.  When a value is null, that is literally the absence of information.  So calculations can run on cells with data, and will simply ignore the null cells.  If there was a string there, it would cause calculations to fail.  That said, there are some functions that allow you do do this in dataflows:

    MySQL / Beastmode -- use the ifnull function

    Redshift -- use the nvl function

    MagicETL -- use a replace values function and select null values in the magicETL.

     

    We are currently using the ifnull method in the beastmode so that we can create a new column (ex. Original column - `Name` -  has Bob, Mike, and unassigned/null values, and create a new column `Names2` which is ifnull(`Names`, 'null')) to account for the blank/null values and have proper filtering. When we don't account for the nulls and try to filter the original column `Names` for "not in Mike," it will also filter out all the rows/columns with blank/unassigned values. Or if we just filter to the blank/unassigned values it can also end up messing up the data. We were wondering if there was a more convenient work around to this instead of having to create new columns to account for nulls/blank values so that our filtering does not get messed up.



    Regarding the tables questions, please provide a little more context on this.  Can you clarify question 1 in general to help me what you are trying to accomplish.

     

    When you have the Chart Type as a Table and go to Chart Properties > General > Hide Columns it has a text field that allows you to hide columns by its index / order in the list of tables but no option to hide by the column variable name itself.



    Regarding question 2, are you using workbench or some other method to import the data in question?


     

    This is a problem we have been encountering regardless of data source. When we have a table and change the label, and then move or delete the column, it can mess up the ordering of the labels and even assign labels to columns that did not have labels .

  • That answers my questions. Thanks Jared!