transforming Data with Pyspark

[omar21]

I have a table which contains 3 million rows and 100 columns ; and I wanted to transform the data with pyspark.

I create a pyspark recipe and I import the data without problem but when I wanted to just execute easy operations like .show or count() it takes a lot of time almost 10min or more

If you have any idea why he does it like that I would be very grateful.

[omar21]

i get the error 

Py4JJavaError: An error occurred while calling o71.count.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 6.0 failed 4 times, most recent failure: Lost task 0.3 in stage 6.0 (TID 10) (10.130.12.246 executor 2): org.apache.spark.SparkException: Error communicating with MapOutputTracker