How does the pyspark mapPartitions function work?

问题

So I am trying to learn Spark using Python (Pyspark). I want to know how the function mapPartitions work. That is what Input it takes and what Output it gives. I couldn't find any proper example from the internet. Lets say, I have an RDD object containing lists, such as below.

[ [1, 2, 3], [3, 2, 4], [5, 2, 7] ] 

And I want to remove element 2 from all the lists, how would I achieve that using mapPartitions.

回答1:

mapPartition should be thought of as a map operation over partitions and not over the elements of the partition. It's input is the set of current partitions its output will be another set of partitions.

The function you pass map must take an individual element of your RDD

The function you pass mapPartition must take an iterable of your RDD type and return and iterable of some other or the same type.

In your case you probably just want to do something like

def filter_out_2(line):
    return [x for x in line if x != 2]

filtered_lists = data.map(filterOut2)

if you wanted to use mapPartition it would be

def filter_out_2_from_partition(list_of_lists):
  final_iterator = []
  for sub_list in list_of_lists:
    final_iterator.append( [x for x in sub_list if x != 2])
  return iter(final_iterator)

filtered_lists = data.mapPartition(filterOut2FromPartion)

回答2:

It's easier to use mapPartitions with a generator function using the yield syntax:

def filter_out_2(partition):
    for element in partition:
        if element != 2:
            yield element

filtered_lists = data.mapPartitions(filter_out_2)

回答3:

Need a final Iter

def filter_out_2(partition):
for element in partition:
    sec_iterator = []
    for i in element:
        if i!= 2:
            sec_iterator.append(i)
    yield sec_iterator

filtered_lists = data.mapPartitions(filter_out_2)
for i in filtered_lists.collect(): print(i)

来源：https://stackoverflow.com/questions/26741714/how-does-the-pyspark-mappartitions-function-work