Hi, Anyone had this problem. I run a test Data aginst my fitted Rpart Object pruned.prediction <- predict( pruned.modeltree, newdata = evaluation, type "prob",na.action= na.omit ) Warning message: 'newdata' had 2019 rows but variable(s) found have 5865 rows My new data: evaluation has 2019 rows but the number of observation in my rpart Object is 5865. ( This is the same number of rows in the DataSet I use to build the rpart Object). I had tried with explicitly removing all "NA" from DataSets but no help. Thanks Sandip