Pyspark partition dataframe for RNN trainning

Question

I have a dataset of flights with one column for aircraft and one for the Flight_ID. Each flight contains an array of 1500 rows and 75 columns corre