Why is Apache Orc RecordReader.searchArgument() not filtering correctly? Why is Apache Orc RecordReader.searchArgument() not filtering correctly? hadoop hadoop