I am working in Azure Synapse Analytics, in PySpark. Say I have a PySpark dataframe df with a partition