NasaLogSparkStreaming - Details for Query 1

Details for Query 1

Submitted Time: 2024/12/11 23:56:58
Duration: 19 s
Succeeded Jobs: 0 1

Show the Stage ID and Task ID that corresponds to the max metric

digraph G { 0 [labelType="html" label=" WriteToDataSourceV2 "]; subgraph cluster1 { isCluster="true"; label="WholeStageCodegen (7)\n \nduration: 57 ms"; 2 [labelType="html" label="Sort sort time: 0 ms peak memory: 64.0 KiB spill size: 0.0 B"]; } 3 [labelType="html" label=" Exchange "]; subgraph cluster4 { isCluster="true"; label="WholeStageCodegen (6)\n \nduration: total (min, med, max (stageId: taskId))\n1.0 s (0 ms, 1 ms, 20 ms (stage 1.0: task 3))"; 5 [labelType="html" label="HashAggregate time in aggregation build total (min, med, max (stageId: taskId)) 0 ms (0 ms, 0 ms, 0 ms (stage 1.0: task 2)) peak memory total (min, med, max (stageId: taskId)) 100.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 1.0: task 2))"]; } 6 [labelType="html" label="StateStoreSave memory used by state total (min, med, max (stageId: taskId)) 81.3 KiB (208.0 B, 208.0 B, 208.0 B (stage 1.0: task 2)) estimated size of state only on current version total (min, med, max (stageId: taskId)) 25.0 KiB (64.0 B, 64.0 B, 64.0 B (stage 1.0: task 2)) time to commit changes total (min, med, max (stageId: taskId)) 16.4 s (26 ms, 29 ms, 114 ms (stage 1.0: task 3)) time to remove total (min, med, max (stageId: taskId)) 0 ms (0 ms, 0 ms, 0 ms (stage 1.0: task 2)) time to update total (min, med, max (stageId: taskId)) 413 ms (0 ms, 1 ms, 18 ms (stage 1.0: task 27))"]; subgraph cluster7 { isCluster="true"; label="WholeStageCodegen (5)\n \nduration: total (min, med, max (stageId: taskId))\n428 ms (0 ms, 1 ms, 18 ms (stage 1.0: task 27))"; 8 [labelType="html" label="HashAggregate time in aggregation build total (min, med, max (stageId: taskId)) 0 ms (0 ms, 0 ms, 0 ms (stage 1.0: task 2)) peak memory total (min, med, max (stageId: taskId)) 100.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 1.0: task 2))"]; } 9 [labelType="html" label=" StateStoreRestore "]; subgraph cluster10 { isCluster="true"; label="WholeStageCodegen (4)\n \nduration: total (min, med, max (stageId: taskId))\n1.7 s (1 ms, 3 ms, 53 ms (stage 1.0: task 3))"; 11 [labelType="html" label="HashAggregate time in aggregation build total (min, med, max (stageId: taskId)) 1 ms (0 ms, 0 ms, 1 ms (stage 1.0: task 3)) peak memory total (min, med, max (stageId: taskId)) 100.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 1.0: task 2))"]; } 12 [labelType="html" label=" Exchange "]; subgraph cluster13 { isCluster="true"; label="WholeStageCodegen (3)\n \nduration: total (min, med, max (stageId: taskId))\n228 ms (84 ms, 144 ms, 144 ms (stage 0.0: task 0))"; 14 [labelType="html" label="HashAggregate time in aggregation build total (min, med, max (stageId: taskId)) 11 ms (4 ms, 7 ms, 7 ms (stage 0.0: task 0)) peak memory total (min, med, max (stageId: taskId)) 512.0 KiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 0.0: task 1))"]; 15 [labelType="html" label=" Project "]; 16 [labelType="html" label=" Filter "]; 17 [labelType="html" label=" Expand "]; 18 [labelType="html" label=" Project "]; } 19 [labelType="html" label=" EventTimeWatermark "]; subgraph cluster20 { isCluster="true"; label="WholeStageCodegen (2)\n \nduration: total (min, med, max (stageId: taskId))\n430 ms (148 ms, 282 ms, 282 ms (stage 0.0: task 0))"; 21 [labelType="html" label=" Project "]; } 22 [labelType="html" label=" Generate "]; subgraph cluster23 { isCluster="true"; label="WholeStageCodegen (1)\n \nduration: total (min, med, max (stageId: taskId))\n633 ms (220 ms, 413 ms, 413 ms (stage 0.0: task 0))"; 24 [labelType="html" label=" Project "]; } 25 [labelType="html" label=" MicroBatchScan "]; 2->0; 3->2; 5->3; 6->5; 8->6; 9->8; 11->9; 12->11; 14->12; 15->14; 16->15; 17->16; 18->17; 19->18; 21->19; 22->21; 24->22; 25->24; }

WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@10942346

Sort [total_bytes#68 DESC NULLS LAST], true, 0

WholeStageCodegen (7)

Exchange rangepartitioning(total_bytes#68 DESC NULLS LAST, 200), true, [id=#103]

HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[sum(cast(bytes#47 as double))])

WholeStageCodegen (6)

StateStoreSave [window#69-T3000ms, hostname#12], state info [ checkpoint = file:/tmp/spark-checkpoints/state, runId = e0c08a0a-5cf3-468b-888d-995173b23fbd, opId = 0, ver = 0, numPartitions = 200], Complete, 0, 2

HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[merge_sum(cast(bytes#47 as double))])

WholeStageCodegen (5)

StateStoreRestore [window#69-T3000ms, hostname#12], state info [ checkpoint = file:/tmp/spark-checkpoints/state, runId = e0c08a0a-5cf3-468b-888d-995173b23fbd, opId = 0, ver = 0, numPartitions = 200], 2

HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[merge_sum(cast(bytes#47 as double))])

WholeStageCodegen (4)

Exchange hashpartitioning(window#69-T3000ms, hostname#12, 200), true, [id=#91]

HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[partial_sum(cast(bytes#47 as double))])

Project [window#69-T3000ms, hostname#12, bytes#47]

Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#69-T3000ms)) AND (timestamp#1-T3000ms >= window#69-T3000ms.start)) AND (timestamp#1-T3000ms < window#69-T3000ms.end))

Expand [ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47), ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47)], [window#69-T3000ms, timestamp#1-T3000ms, hostname#12, bytes#47]

Project [timestamp#1-T3000ms, split(logs#5, ,, -1)[1] AS hostname#12, split(logs#5, ,, -1)[6] AS bytes#47]

WholeStageCodegen (3)

EventTimeWatermark timestamp#1: timestamp, 3 seconds

Project [logs#5, timestamp#1]

WholeStageCodegen (2)

Generate explode(split(value#0, , -1)), [timestamp#1], false, [logs#5]

Project [value#0, timestamp#1]

WholeStageCodegen (1)

MicroBatchScan[value#0, timestamp#1] class org.apache.spark.sql.execution.streaming.sources.TextSocketTable$$anon$1

Details

== Parsed Logical Plan ==
WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@10942346
+- Sort [total_bytes#68 DESC NULLS LAST], true
   +- Aggregate [window#69-T3000ms, hostname#12], [window#69-T3000ms AS window#57-T3000ms, hostname#12, sum(cast(bytes#47 as double)) AS total_bytes#68]
      +- Filter ((timestamp#1-T3000ms >= window#69-T3000ms.start) AND (timestamp#1-T3000ms < window#69-T3000ms.end))
         +- Expand [ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(0 as bigint)) - cast(2 as bigint)) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion((((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(0 as bigint)) - cast(2 as bigint)) * 30000000) + 0) + 60000000), LongType, TimestampType)), logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47), ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(1 as bigint)) - cast(2 as bigint)) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion((((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(1 as bigint)) - cast(2 as bigint)) * 30000000) + 0) + 60000000), LongType, TimestampType)), logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47)], [window#69-T3000ms, logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47]
            +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, split(logs#5, ,, -1)[6] AS bytes#47]
               +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, split(logs#5, ,, -1)[5] AS responsecode#38]
                  +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, split(logs#5, ,, -1)[4] AS url#30]
                     +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, split(logs#5, ,, -1)[3] AS method#23]
                        +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, split(logs#5, ,, -1)[2] AS time#17]
                           +- Project [logs#5, timestamp#1-T3000ms, idx#8, split(logs#5, ,, -1)[1] AS hostname#12]
                              +- Project [logs#5, timestamp#1-T3000ms, split(logs#5, ,, -1)[0] AS idx#8]
                                 +- EventTimeWatermark timestamp#1: timestamp, 3 seconds
                                    +- Project [logs#5, timestamp#1]
                                       +- Generate explode(split(value#0,  , -1)), false, [logs#5]
                                          +- StreamingDataSourceV2Relation [value#0, timestamp#1], org.apache.spark.sql.execution.streaming.sources.TextSocketTable$$anon$1@64e6357e, TextSocketV2[host: stream-emulator.data-science-tools.svc.cluster.local, port: 5551], -1, -1

== Analyzed Logical Plan ==

WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@10942346
+- Sort [total_bytes#68 DESC NULLS LAST], true
   +- Aggregate [window#69-T3000ms, hostname#12], [window#69-T3000ms AS window#57-T3000ms, hostname#12, sum(cast(bytes#47 as double)) AS total_bytes#68]
      +- Filter ((timestamp#1-T3000ms >= window#69-T3000ms.start) AND (timestamp#1-T3000ms < window#69-T3000ms.end))
         +- Expand [ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(0 as bigint)) - cast(2 as bigint)) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion((((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(0 as bigint)) - cast(2 as bigint)) * 30000000) + 0) + 60000000), LongType, TimestampType)), logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47), ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(1 as bigint)) - cast(2 as bigint)) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion((((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) + cast(1 as bigint)) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / cast(30000000 as double))) END + cast(1 as bigint)) - cast(2 as bigint)) * 30000000) + 0) + 60000000), LongType, TimestampType)), logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47)], [window#69-T3000ms, logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, bytes#47]
            +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, responsecode#38, split(logs#5, ,, -1)[6] AS bytes#47]
               +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, url#30, split(logs#5, ,, -1)[5] AS responsecode#38]
                  +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, method#23, split(logs#5, ,, -1)[4] AS url#30]
                     +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, time#17, split(logs#5, ,, -1)[3] AS method#23]
                        +- Project [logs#5, timestamp#1-T3000ms, idx#8, hostname#12, split(logs#5, ,, -1)[2] AS time#17]
                           +- Project [logs#5, timestamp#1-T3000ms, idx#8, split(logs#5, ,, -1)[1] AS hostname#12]
                              +- Project [logs#5, timestamp#1-T3000ms, split(logs#5, ,, -1)[0] AS idx#8]
                                 +- EventTimeWatermark timestamp#1: timestamp, 3 seconds
                                    +- Project [logs#5, timestamp#1]
                                       +- Generate explode(split(value#0,  , -1)), false, [logs#5]
                                          +- StreamingDataSourceV2Relation [value#0, timestamp#1], org.apache.spark.sql.execution.streaming.sources.TextSocketTable$$anon$1@64e6357e, TextSocketV2[host: stream-emulator.data-science-tools.svc.cluster.local, port: 5551], -1, -1

== Optimized Logical Plan ==
WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@10942346
+- Sort [total_bytes#68 DESC NULLS LAST], true
   +- Aggregate [window#69-T3000ms, hostname#12], [window#69-T3000ms AS window#57-T3000ms, hostname#12, sum(cast(bytes#47 as double)) AS total_bytes#68]
      +- Project [window#69-T3000ms, hostname#12, bytes#47]
         +- Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#69-T3000ms)) AND (timestamp#1-T3000ms >= window#69-T3000ms.start)) AND (timestamp#1-T3000ms < window#69-T3000ms.end))
            +- Expand [ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47), ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47)], [window#69-T3000ms, timestamp#1-T3000ms, hostname#12, bytes#47]
               +- Project [timestamp#1-T3000ms, split(logs#5, ,, -1)[1] AS hostname#12, split(logs#5, ,, -1)[6] AS bytes#47]
                  +- EventTimeWatermark timestamp#1: timestamp, 3 seconds
                     +- Project [logs#5, timestamp#1]
                        +- Generate explode(split(value#0,  , -1)), [0], false, [logs#5]
                           +- StreamingDataSourceV2Relation [value#0, timestamp#1], org.apache.spark.sql.execution.streaming.sources.TextSocketTable$$anon$1@64e6357e, TextSocketV2[host: stream-emulator.data-science-tools.svc.cluster.local, port: 5551], -1, -1

== Physical Plan ==
WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@10942346
+- *(7) Sort [total_bytes#68 DESC NULLS LAST], true, 0
   +- Exchange rangepartitioning(total_bytes#68 DESC NULLS LAST, 200), true, [id=#103]
      +- *(6) HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[sum(cast(bytes#47 as double))], output=[window#57-T3000ms, hostname#12, total_bytes#68])
         +- StateStoreSave [window#69-T3000ms, hostname#12], state info [ checkpoint = file:/tmp/spark-checkpoints/state, runId = e0c08a0a-5cf3-468b-888d-995173b23fbd, opId = 0, ver = 0, numPartitions = 200], Complete, 0, 2
            +- *(5) HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[merge_sum(cast(bytes#47 as double))], output=[window#69-T3000ms, hostname#12, sum#77])
               +- StateStoreRestore [window#69-T3000ms, hostname#12], state info [ checkpoint = file:/tmp/spark-checkpoints/state, runId = e0c08a0a-5cf3-468b-888d-995173b23fbd, opId = 0, ver = 0, numPartitions = 200], 2
                  +- *(4) HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[merge_sum(cast(bytes#47 as double))], output=[window#69-T3000ms, hostname#12, sum#77])
                     +- Exchange hashpartitioning(window#69-T3000ms, hostname#12, 200), true, [id=#91]
                        +- *(3) HashAggregate(keys=[window#69-T3000ms, hostname#12], functions=[partial_sum(cast(bytes#47 as double))], output=[window#69-T3000ms, hostname#12, sum#77])
                           +- *(3) Project [window#69-T3000ms, hostname#12, bytes#47]
                              +- *(3) Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#69-T3000ms)) AND (timestamp#1-T3000ms >= window#69-T3000ms.start)) AND (timestamp#1-T3000ms < window#69-T3000ms.end))
                                 +- *(3) Expand [ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 0) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47), ArrayBuffer(named_struct(start, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 0), LongType, TimestampType), end, precisetimestampconversion(((((CASE WHEN (cast(CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) as double) = (cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) THEN (CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) + 1) ELSE CEIL((cast((precisetimestampconversion(timestamp#1-T3000ms, TimestampType, LongType) - 0) as double) / 3.0E7)) END + 1) - 2) * 30000000) + 60000000), LongType, TimestampType)), timestamp#1-T3000ms, hostname#12, bytes#47)], [window#69-T3000ms, timestamp#1-T3000ms, hostname#12, bytes#47]
                                    +- *(3) Project [timestamp#1-T3000ms, split(logs#5, ,, -1)[1] AS hostname#12, split(logs#5, ,, -1)[6] AS bytes#47]
                                       +- EventTimeWatermark timestamp#1: timestamp, 3 seconds
                                          +- *(2) Project [logs#5, timestamp#1]
                                             +- Generate explode(split(value#0,  , -1)), [timestamp#1], false, [logs#5]
                                                +- *(1) Project [value#0, timestamp#1]
                                                   +- MicroBatchScan[value#0, timestamp#1] class org.apache.spark.sql.execution.streaming.sources.TextSocketTable$$anon$1