Pyspark column is nan. alias(c) for c in data.