Uploaded image for project: 'Apache MADlib'
  1. Apache MADlib
  2. MADLIB-879

Random Forest function gives error for some expressions.

Attach filesAttach ScreenshotAdd voteVotersWatch issueWatchersCreate sub-taskLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments


    • Bug
    • Status: Resolved
    • Major
    • Resolution: Unresolved
    • None
    • None
    • None


      I have been trying to make the new random forest function of MADlib 1.7 but I get errors in one version of the function definition, but not with another.

      I tried a few things and got errors when I reduced the arguments from 14 to mandatory ones only but got the plpy error as shown in SQL4_result.png. Let me explain what I experienced below.

      SQL3: I used 14 arguments and it worked (please see SQL3.png as attached). However, when I changed the list_of_features to be '*', I got the plpy error. Then, it worked again when I changed the list-of-variables-to-exclude from NULL to 'id'. The version of this function is shown in SQL3_version.png.

      SQL4: Removed the highlighted arguments (please see SQL4.png) and now I have arguments that are mandatory for this function. I got the plpy error this time. I could not fix this error by changing the list of features to be *, or changing the list-of-variables-to-exclude from NULL to 'id'. The version of this SQL is shown in SQL4_version.png. I do not see any problems in setting arguments.

      What I can guess from here is that the NULL value for the list-of-variables-to-exclude may not be the only issue because I got the plpy error by just reducing the number of arguments. The values of arguments in SQL3 are all set as default values and therefore reducing those optional arguments should not affect the execution of the function.

      Do you see any problems in my SQL or arguments?



          This comment will be Viewable by All Users Viewable by All Users


            pjayaram Preethi Jayaram
            tkobayashi Tetsuo Kobayashi




                Issue deployment