@sahil07, the fact that you could read a csv file using spark.read.csv() is because you're using the spark native API to access the dbfs, which works just fine. The PDF reading was not possible because PyPDF2 does not use the spark API, but python standard reader.