digraph G {
0 [labelType="html" label="<br><b>WriteToDataSourceV2</b><br><br>"];
subgraph cluster1 {
isCluster="true";
label="WholeStageCodegen (6)\n \nduration: total (min, med, max (stageId: taskId))\n11.5 s (37 ms, 45 ms, 170 ms (stage 21.0: task 42))";
2 [labelType="html" label="<b>HashAggregate</b><br><br>time in aggregation build total (min, med, max (stageId: taskId))<br>11.1 s (36 ms, 43 ms, 162 ms (stage 21.0: task 42))<br>peak memory total (min, med, max (stageId: taskId))<br>50.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 21.0: task 43))"];
}
3 [labelType="html" label="<b>StateStoreSave</b><br><br>memory used by state total (min, med, max (stageId: taskId))<br>43.8 KiB (224.0 B, 224.0 B, 224.0 B (stage 21.0: task 43))<br>estimated size of state only on current version total (min, med, max (stageId: taskId))<br>15.6 KiB (80.0 B, 80.0 B, 80.0 B (stage 21.0: task 43))<br>time to commit changes total (min, med, max (stageId: taskId))<br>10.4 s (34 ms, 40 ms, 141 ms (stage 21.0: task 42))<br>time to remove total (min, med, max (stageId: taskId))<br>63 ms (0 ms, 0 ms, 6 ms (stage 21.0: task 42))<br>time to update total (min, med, max (stageId: taskId))<br>710 ms (1 ms, 3 ms, 42 ms (stage 21.0: task 42))"];
subgraph cluster4 {
isCluster="true";
label="WholeStageCodegen (5)\n \nduration: total (min, med, max (stageId: taskId))\n938 ms (2 ms, 4 ms, 73 ms (stage 21.0: task 42))";
5 [labelType="html" label="<b>HashAggregate</b><br><br>time in aggregation build total (min, med, max (stageId: taskId))<br>5 ms (0 ms, 0 ms, 4 ms (stage 21.0: task 106))<br>peak memory total (min, med, max (stageId: taskId))<br>50.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 21.0: task 43))"];
}
6 [labelType="html" label="<br><b>StateStoreRestore</b><br><br>"];
subgraph cluster7 {
isCluster="true";
label="WholeStageCodegen (4)\n \nduration: total (min, med, max (stageId: taskId))\n1.9 s (4 ms, 8 ms, 132 ms (stage 21.0: task 42))";
8 [labelType="html" label="<b>HashAggregate</b><br><br>time in aggregation build total (min, med, max (stageId: taskId))<br>4 ms (0 ms, 0 ms, 3 ms (stage 21.0: task 42))<br>peak memory total (min, med, max (stageId: taskId))<br>50.0 MiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 21.0: task 43))"];
}
9 [labelType="html" label="<br><b>Exchange</b><br><br>"];
subgraph cluster10 {
isCluster="true";
label="WholeStageCodegen (3)\n \nduration: total (min, med, max (stageId: taskId))\n223 ms (79 ms, 144 ms, 144 ms (stage 20.0: task 40))";
11 [labelType="html" label="<b>HashAggregate</b><br><br>time in aggregation build total (min, med, max (stageId: taskId))<br>10 ms (4 ms, 6 ms, 6 ms (stage 20.0: task 40))<br>peak memory total (min, med, max (stageId: taskId))<br>512.0 KiB (256.0 KiB, 256.0 KiB, 256.0 KiB (stage 20.0: task 41))"];
12 [labelType="html" label="<br><b>Project</b><br><br>"];
13 [labelType="html" label="<br><b>Filter</b><br><br>"];
14 [labelType="html" label="<br><b>Expand</b><br><br>"];
}
15 [labelType="html" label="<br><b>EventTimeWatermark</b><br><br>"];
subgraph cluster16 {
isCluster="true";
label="WholeStageCodegen (2)\n \nduration: total (min, med, max (stageId: taskId))\n424 ms (163 ms, 261 ms, 261 ms (stage 20.0: task 40))";
17 [labelType="html" label="<br><b>Project</b><br><br>"];
18 [labelType="html" label="<br><b>Filter</b><br><br>"];
}
19 [labelType="html" label="<br><b>Generate</b><br><br>"];
subgraph cluster20 {
isCluster="true";
label="WholeStageCodegen (1)\n \nduration: total (min, med, max (stageId: taskId))\n469 ms (180 ms, 289 ms, 289 ms (stage 20.0: task 40))";
21 [labelType="html" label="<br><b>Project</b><br><br>"];
}
22 [labelType="html" label="<br><b>MicroBatchScan</b><br><br>"];
2->0;
3->2;
5->3;
6->5;
8->6;
9->8;
11->9;
12->11;
13->12;
14->13;
15->14;
17->15;
18->17;
19->18;
21->19;
22->21;
}
23
WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@3e8c9c4c
HashAggregate(keys=[window#822-T3000ms], functions=[count(1)])
WholeStageCodegen (6)
StateStoreSave [window#822-T3000ms], state info [ checkpoint = file:/tmp/temporary-a4a518b2-a3ff-4163-8a29-c5a18d39c201/state, runId = 22115180-3cbd-43f4-9c89-7ad00fbe0238, opId = 0, ver = 0, numPartitions = 200], Update, 0, 2
HashAggregate(keys=[window#822-T3000ms], functions=[merge_count(1)])
WholeStageCodegen (5)
StateStoreRestore [window#822-T3000ms], state info [ checkpoint = file:/tmp/temporary-a4a518b2-a3ff-4163-8a29-c5a18d39c201/state, runId = 22115180-3cbd-43f4-9c89-7ad00fbe0238, opId = 0, ver = 0, numPartitions = 200], 2
HashAggregate(keys=[window#822-T3000ms], functions=[merge_count(1)])
WholeStageCodegen (4)
Exchange hashpartitioning(window#822-T3000ms, 200), true, [id=#775]
HashAggregate(keys=[window#822-T3000ms], functions=[partial_count(1)])
Project [window#822-T3000ms]
Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#822-T3000ms)) AND (timestamp#1-T3000ms >= window#822-T3000ms.start)) AND (timestamp#1-T3000ms < window#822-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), 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)], [window#822-T3000ms, timestamp#1-T3000ms]
WholeStageCodegen (3)
EventTimeWatermark timestamp#1: timestamp, 3 seconds
Project [timestamp#1]
Filter Contains(split(logs#5, ,, -1)[4], gif)
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
== Parsed Logical Plan ==
WriteToDataSourceV2 org.apache.spark.sql.execution.streaming.sources.MicroBatchWrite@3e8c9c4c
+- Project [window#810-T3000ms, count#821L AS gif_count#825L]
+- Aggregate [window#822-T3000ms], [window#822-T3000ms AS window#810-T3000ms, count(1) AS count#821L]
+- Filter ((timestamp#1-T3000ms >= window#822-T3000ms.start) AND (timestamp#1-T3000ms < window#822-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#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175), 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#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175)], [window#822-T3000ms, logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175]
+- Filter url#158 LIKE %gif%
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, responsecode#166, split(logs#5, ,, -1)[6] AS bytes#175]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, split(logs#5, ,, -1)[5] AS responsecode#166]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, split(logs#5, ,, -1)[4] AS url#158]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, split(logs#5, ,, -1)[3] AS method#151]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, split(logs#5, ,, -1)[2] AS time#145]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, split(logs#5, ,, -1)[1] AS hostname#140]
+- Project [logs#5, timestamp#1-T3000ms, split(logs#5, ,, -1)[0] AS idx#136]
+- 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@7eafcdef, 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@3e8c9c4c
+- Project [window#810-T3000ms, count#821L AS gif_count#825L]
+- Aggregate [window#822-T3000ms], [window#822-T3000ms AS window#810-T3000ms, count(1) AS count#821L]
+- Filter ((timestamp#1-T3000ms >= window#822-T3000ms.start) AND (timestamp#1-T3000ms < window#822-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#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175), 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#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175)], [window#822-T3000ms, logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, responsecode#166, bytes#175]
+- Filter url#158 LIKE %gif%
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, responsecode#166, split(logs#5, ,, -1)[6] AS bytes#175]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, url#158, split(logs#5, ,, -1)[5] AS responsecode#166]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, method#151, split(logs#5, ,, -1)[4] AS url#158]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, time#145, split(logs#5, ,, -1)[3] AS method#151]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, hostname#140, split(logs#5, ,, -1)[2] AS time#145]
+- Project [logs#5, timestamp#1-T3000ms, idx#136, split(logs#5, ,, -1)[1] AS hostname#140]
+- Project [logs#5, timestamp#1-T3000ms, split(logs#5, ,, -1)[0] AS idx#136]
+- 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@7eafcdef, 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@3e8c9c4c
+- Aggregate [window#822-T3000ms], [window#822-T3000ms AS window#810-T3000ms, count(1) AS gif_count#825L]
+- Project [window#822-T3000ms]
+- Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#822-T3000ms)) AND (timestamp#1-T3000ms >= window#822-T3000ms.start)) AND (timestamp#1-T3000ms < window#822-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), 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)], [window#822-T3000ms, timestamp#1-T3000ms]
+- EventTimeWatermark timestamp#1: timestamp, 3 seconds
+- Project [timestamp#1]
+- Filter Contains(split(logs#5, ,, -1)[4], gif)
+- 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@7eafcdef, 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@3e8c9c4c
+- *(6) HashAggregate(keys=[window#822-T3000ms], functions=[count(1)], output=[window#810-T3000ms, gif_count#825L])
+- StateStoreSave [window#822-T3000ms], state info [ checkpoint = file:/tmp/temporary-a4a518b2-a3ff-4163-8a29-c5a18d39c201/state, runId = 22115180-3cbd-43f4-9c89-7ad00fbe0238, opId = 0, ver = 0, numPartitions = 200], Update, 0, 2
+- *(5) HashAggregate(keys=[window#822-T3000ms], functions=[merge_count(1)], output=[window#822-T3000ms, count#831L])
+- StateStoreRestore [window#822-T3000ms], state info [ checkpoint = file:/tmp/temporary-a4a518b2-a3ff-4163-8a29-c5a18d39c201/state, runId = 22115180-3cbd-43f4-9c89-7ad00fbe0238, opId = 0, ver = 0, numPartitions = 200], 2
+- *(4) HashAggregate(keys=[window#822-T3000ms], functions=[merge_count(1)], output=[window#822-T3000ms, count#831L])
+- Exchange hashpartitioning(window#822-T3000ms, 200), true, [id=#775]
+- *(3) HashAggregate(keys=[window#822-T3000ms], functions=[partial_count(1)], output=[window#822-T3000ms, count#831L])
+- *(3) Project [window#822-T3000ms]
+- *(3) Filter (((isnotnull(timestamp#1-T3000ms) AND isnotnull(window#822-T3000ms)) AND (timestamp#1-T3000ms >= window#822-T3000ms.start)) AND (timestamp#1-T3000ms < window#822-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), 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)], [window#822-T3000ms, timestamp#1-T3000ms]
+- EventTimeWatermark timestamp#1: timestamp, 3 seconds
+- *(2) Project [timestamp#1]
+- *(2) Filter Contains(split(logs#5, ,, -1)[4], gif)
+- 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