BryanCutler commented on issue #24070: [SPARK-23961][PYTHON] Fix error when toLocalIterator goes out of scope URL: https://github.com/apache/spark/pull/24070#issuecomment-472129446 ### Timings for DataFrame.toLocalIterator and RDD.toLocalIterator These tests are to illustrate the slowdown caused by this change, comparing current master with this change. Wall clock time is measure to fully consume the local iterator and average of 5 runs are shown: _ | DataFrame | RDD --- | --- | --- *master* | 10.26016583 | 4.354181528 *this PR* | 12.14033799 | 3.823320436 #### Test Script ```python import time from pyspark.sql import SparkSession spark = SparkSession\ .builder\ .appName("toLocalIterator_timing")\ .getOrCreate() num = 1 << 22 numParts = 32 def run(df): print("Starting iterator:") start = time.time() count = 0 for row in df.toLocalIterator(): count += 1 if count != num: raise RuntimeError("Expected {} but got {}".format(num, count)) elapsed = time.time() - start print("completed in {}".format(elapsed)) run(spark.range(num, numPartitions=numParts)) run(spark.sparkContext.range(num, numSlices=numParts)) spark.stop() ```
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
