More actions
Fondamenti di Big Data con PySpark
Upendra Devisetty
Science Analyst, CyVerse
reduce() action
reduce(func) action is used for aggregating the elements of a regular RDD
The function should be commutative (changing the order of the operands does not change the result) and associative
An example of
reduce()action in PySpark
x = [1,3,4,6]
RDD = sc.parallelize(x)
RDD.reduce(lambda x, y : x + y)
14
saveAsTextFile() action
saveAsTextFile()action saves RDD into a text file inside a directory with each partition as a separate file
RDD.saveAsTextFile("tempFile")
coalesce()method can be used to save RDD as a single text file
RDD.coalesce(1).saveAsTextFile("tempFile")
Action Operations on pair RDDs
RDD actions available for PySpark pair RDDs
Pair RDD actions leverage the key-value data
Few examples of pair RDD actions include
countByKey()collectAsMap()
countByKey() action
countByKey()only available for type (K, V)countByKey()action counts the number of elements for each keyExample of
countByKey()on a simple list
rdd = sc.parallelize([("a", 1), ("b", 1), ("a", 1)])
for kee, val in rdd.countByKey().items():
print(kee, val)
('a', 2)
('b', 1)
collectAsMap() action
collectAsMap()return the key-value pairs in the RDD as a dictionaryExample of
collectAsMap()on a simple tuple
sc.parallelize([(1, 2), (3, 4)]).collectAsMap()
{1: 2, 3: 4}
Let's practice
Fondamenti di Big Data con PySpark
