Replacing Null Fields with Imputation

[ROUD89]

Hi,

I have an error using imputation in a dataset. When I use imputation I choose to replace [NULL] with the median. However, the result is that EVERY value in this variable becomes [NULL] instead of the desired effect. Any help here?

These are my variables, the variable that I wan to impute is Age-years, with field type 'Double' 

As you can see in my photos, the Age-years variable has become solely [NULL] values.my variablesmy imputationthe result which is not desired

[SeanAdams · Accepted solution]

Hey @ROUD89

The imputation tool does do funny things with field names - I did the same as in your example, and the field with the original name comes back all null, but a new field with good values but a slightly different name (instead of age-years, it comes back as age_years) comes back with correct values.

I suspect that the reason that the imputation control does this is because it depends on the cross-tab control under the covers (which has this odd behavior with field names too).

Have a look at the attached workflow - hope this helps.    If not - feel free to post an alteryx workflow with a test data set and we can work together as a community to crack it with you.

Cheers

Sean

[JeffF]

This issue is resolved in the 2020.2 release available here: http://downloads.alteryx.com

The latest version of the Imputation tool (v3) is used when adding the tool to the canvas in 2020.2.

To update an Imputation tool in an existing workflow, right-click the tool and select Imputation 3.

[TimothyL]

Hi @JeffF @ROUD89 @SeanAdams, 

We have built some new missing value imputation macros here: https://community.alteryx.com/t5/Data-Science/Expand-Your-Predictive-Palette-IV-Imputation-Beyond-Me...

Now everyone could use other imputation methods to replace Nulls. Give it a try!

TL