The Rapid Deployment node is used to deploy PMML models on the testing data. The spreadsheet below shows an example of a summary table out of the Rapid Deployment Node. The first column displays the observed dependent variable, the second column displays the Predicted dependent value and the third column displays the Residual value. You may notice that there are missing values across some rows of the table. A missing value in the first column indicates the response variable for that row is missing in the testing data. While a missing value in the second column usually indicates that missing values are found among predictors/independent variables within those samples, which makes the prediction result unavailable. As a result,  the computation for the residual(the third column) can not be obtained.

This article talks about missing values in the output table of the Rapid Deployment node.

For Tree models, it is possible to impute missing predictors for calculating predictions. This option can be enabled by checking "Predict case(s) with missing data in inputs" under the Rapid Deployment node.