Window function requires a timestamp column because you will apply a function for each window (like an aggregation). You still can use UDF for customized tasks
Em 25 de dez de 2017 20:15, "M Singh" <mans2si...@yahoo.com.invalid> escreveu: > Hi: > I would like to use window function on a DataSet stream (Spark 2.2.0) > The window function requires Column as argument and can be used with > DataFrames by passing the column. Is there any analogous window function or > pointers to how window function can be used for DataSets ? > > Thanks >
