Details
Description
The following Spark shell snippet reproduces this issue:
import org.apache.spark.sql.functions._ val parallelism = 8 // Adjust this to default parallelism val df = sqlContext. range(2 * parallelism). // 8 partitions, 2 elements per partition select( col("id"), monotonicallyIncreasingId().as("long_id") ) df.show() // +---+-----------+ // | id| long_id| // +---+-----------+ // | 0| 0| // | 1| 1| // | 2| 8589934592| // | 3| 8589934593| // | 4|17179869184| // | 5|17179869185| // | 6|25769803776| // | 7|25769803777| // | 8|34359738368| // | 9|34359738369| // | 10|42949672960| // | 11|42949672961| // | 12|51539607552| // | 13|51539607553| // | 14|60129542144| // | 15|60129542145| // +---+-----------+ df. filter(col("id") === 3). // 2nd element in the 2nd partition show() // +---+----------+ // | id| long_id| // +---+----------+ // | 3|8589934592| // +---+----------+
monotonicallyIncreasingId is nondeterministic.
Attachments
Issue Links
- is related to
-
SPARK-14241 Output of monotonically_increasing_id lacks stable relation with rows of DataFrame
-
- Resolved
-
- links to
