how to solve this : TypeError: expected string or bytes-like object

Question

I am trying to do tokenize my text. The type of my data is pyspark.sql.dataframe.DataFrame I read it by the following code:

df1=spark.read.json(&qu