Why does column 1st_from_end
contain null:
from pyspark.sql.functions import split
df = sqlContext.createDataFrame([(\'a b c d\',)], [\'s\',])
d
If you're using Spark >= 2.4.0 see jxc's answer below.
In Spark < 2.4.0, dataframes API didn't support -1
indexing on arrays in Spark < 2.4.0, but you could write own UDF or use built-in size()
function, for example:
>>> from pyspark.sql.functions import size
>>> splitted = df.select(split(df.s, ' ').alias('arr'))
>>> splitted.select(splitted.arr[size(splitted.arr)-1]).show()
+--------------------+
|arr[(size(arr) - 1)]|
+--------------------+
| d|
+--------------------+
Building on jamiet 's solution, we can simplify even further by removing a reverse
from pyspark.sql.functions import split, reverse
df = sqlContext.createDataFrame([('a b c d',)], ['s',])
df.select( split(df.s, ' ')[0].alias('0th'),
split(df.s, ' ')[3].alias('3rd'),
reverse(split(df.s, ' '))[-1].alias('1st_from_end')
).show()
Create your own udf would look like this
def get_last_element(l):
return l[-1]
get_last_element_udf = F.udf(get_last_element)
df.select(get_last_element(split(df.s, ' ')).alias('1st_from_end')
For Spark 2.4+, use pyspark.sql.functions.element_at, see below from the documentation:
element_at(array, index) - Returns element of array at given (1-based) index. If index < 0, accesses elements from the last to the first. Returns NULL if the index exceeds the length of the array.
from pyspark.sql.functions import element_at, split, col
df = spark.createDataFrame([('a b c d',)], ['s',])
df.withColumn('arr', split(df.s, ' ')) \
.select( col('arr')[0].alias('0th')
, col('arr')[3].alias('3rd')
, element_at(col('arr'), -1).alias('1st_from_end')
).show()
+---+---+------------+
|0th|3rd|1st_from_end|
+---+---+------------+
| a| d| d|
+---+---+------------+