/
SparkStrategies.scala
519 lines (456 loc) · 23.6 KB
/
SparkStrategies.scala
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.spark.sql.execution
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.{execution, AnalysisException, Strategy}
import org.apache.spark.sql.catalyst.InternalRow
import org.apache.spark.sql.catalyst.encoders.RowEncoder
import org.apache.spark.sql.catalyst.expressions._
import org.apache.spark.sql.catalyst.planning._
import org.apache.spark.sql.catalyst.plans._
import org.apache.spark.sql.catalyst.plans.logical._
import org.apache.spark.sql.catalyst.plans.physical._
import org.apache.spark.sql.execution.columnar.{InMemoryRelation, InMemoryTableScanExec}
import org.apache.spark.sql.execution.command._
import org.apache.spark.sql.execution.exchange.ShuffleExchangeExec
import org.apache.spark.sql.execution.joins.{BuildLeft, BuildRight, BuildSide}
import org.apache.spark.sql.execution.streaming._
import org.apache.spark.sql.execution.streaming.sources.MemoryPlanV2
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.streaming.StreamingQuery
import org.apache.spark.sql.types.StructType
/**
* Converts a logical plan into zero or more SparkPlans. This API is exposed for experimenting
* with the query planner and is not designed to be stable across spark releases. Developers
* writing libraries should instead consider using the stable APIs provided in
* [[org.apache.spark.sql.sources]]
*/
abstract class SparkStrategy extends GenericStrategy[SparkPlan] {
override protected def planLater(plan: LogicalPlan): SparkPlan = PlanLater(plan)
}
case class PlanLater(plan: LogicalPlan) extends LeafExecNode {
override def output: Seq[Attribute] = plan.output
protected override def doExecute(): RDD[InternalRow] = {
throw new UnsupportedOperationException()
}
}
abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
self: SparkPlanner =>
/**
* Plans special cases of limit operators.
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
case Limit(IntegerLiteral(limit), Sort(order, true, child)) =>
TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child))) =>
TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
case Limit(IntegerLiteral(limit), child) =>
// With whole stage codegen, Spark releases resources only when all the output data of the
// query plan are consumed. It's possible that `CollectLimitExec` only consumes a little
// data from child plan and finishes the query without releasing resources. Here we wrap
// the child plan with `LocalLimitExec`, to stop the processing of whole stage codegen and
// trigger the resource releasing work, after we consume `limit` rows.
CollectLimitExec(limit, LocalLimitExec(limit, planLater(child))) :: Nil
case other => planLater(other) :: Nil
}
case Limit(IntegerLiteral(limit), Sort(order, true, child)) =>
TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child))) =>
TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
case _ => Nil
}
}
/**
* Select the proper physical plan for join based on joining keys and size of logical plan.
*
* At first, uses the [[ExtractEquiJoinKeys]] pattern to find joins where at least some of the
* predicates can be evaluated by matching join keys. If found, Join implementations are chosen
* with the following precedence:
*
* - Broadcast: We prefer to broadcast the join side with an explicit broadcast hint(e.g. the
* user applied the [[org.apache.spark.sql.functions.broadcast()]] function to a DataFrame).
* If both sides have the broadcast hint, we prefer to broadcast the side with a smaller
* estimated physical size. If neither one of the sides has the broadcast hint,
* we only broadcast the join side if its estimated physical size that is smaller than
* the user-configurable [[SQLConf.AUTO_BROADCASTJOIN_THRESHOLD]] threshold.
* - Shuffle hash join: if the average size of a single partition is small enough to build a hash
* table.
* - Sort merge: if the matching join keys are sortable.
*
* If there is no joining keys, Join implementations are chosen with the following precedence:
* - BroadcastNestedLoopJoin: if one side of the join could be broadcasted
* - CartesianProduct: for Inner join
* - BroadcastNestedLoopJoin
*/
object JoinSelection extends Strategy with PredicateHelper {
/**
* Matches a plan whose output should be small enough to be used in broadcast join.
*/
private def canBroadcast(plan: LogicalPlan): Boolean = {
plan.stats.sizeInBytes >= 0 && plan.stats.sizeInBytes <= conf.autoBroadcastJoinThreshold
}
/**
* Matches a plan whose single partition should be small enough to build a hash table.
*
* Note: this assume that the number of partition is fixed, requires additional work if it's
* dynamic.
*/
private def canBuildLocalHashMap(plan: LogicalPlan): Boolean = {
plan.stats.sizeInBytes < conf.autoBroadcastJoinThreshold * conf.numShufflePartitions
}
/**
* Returns whether plan a is much smaller (3X) than plan b.
*
* The cost to build hash map is higher than sorting, we should only build hash map on a table
* that is much smaller than other one. Since we does not have the statistic for number of rows,
* use the size of bytes here as estimation.
*/
private def muchSmaller(a: LogicalPlan, b: LogicalPlan): Boolean = {
a.stats.sizeInBytes * 3 <= b.stats.sizeInBytes
}
private def canBuildRight(joinType: JoinType): Boolean = joinType match {
case _: InnerLike | LeftOuter | LeftSemi | LeftAnti => true
case j: ExistenceJoin => true
case _ => false
}
private def canBuildLeft(joinType: JoinType): Boolean = joinType match {
case _: InnerLike | RightOuter => true
case _ => false
}
private def broadcastSide(
canBuildLeft: Boolean,
canBuildRight: Boolean,
left: LogicalPlan,
right: LogicalPlan): BuildSide = {
def smallerSide =
if (right.stats.sizeInBytes <= left.stats.sizeInBytes) BuildRight else BuildLeft
if (canBuildRight && canBuildLeft) {
// Broadcast smaller side base on its estimated physical size
// if both sides have broadcast hint
smallerSide
} else if (canBuildRight) {
BuildRight
} else if (canBuildLeft) {
BuildLeft
} else {
// for the last default broadcast nested loop join
smallerSide
}
}
private def canBroadcastByHints(joinType: JoinType, left: LogicalPlan, right: LogicalPlan)
: Boolean = {
val buildLeft = canBuildLeft(joinType) && left.stats.hints.broadcast
val buildRight = canBuildRight(joinType) && right.stats.hints.broadcast
buildLeft || buildRight
}
private def broadcastSideByHints(joinType: JoinType, left: LogicalPlan, right: LogicalPlan)
: BuildSide = {
val buildLeft = canBuildLeft(joinType) && left.stats.hints.broadcast
val buildRight = canBuildRight(joinType) && right.stats.hints.broadcast
broadcastSide(buildLeft, buildRight, left, right)
}
private def canBroadcastBySizes(joinType: JoinType, left: LogicalPlan, right: LogicalPlan)
: Boolean = {
val buildLeft = canBuildLeft(joinType) && canBroadcast(left)
val buildRight = canBuildRight(joinType) && canBroadcast(right)
buildLeft || buildRight
}
private def broadcastSideBySizes(joinType: JoinType, left: LogicalPlan, right: LogicalPlan)
: BuildSide = {
val buildLeft = canBuildLeft(joinType) && canBroadcast(left)
val buildRight = canBuildRight(joinType) && canBroadcast(right)
broadcastSide(buildLeft, buildRight, left, right)
}
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
// --- BroadcastHashJoin --------------------------------------------------------------------
// broadcast hints were specified
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if canBroadcastByHints(joinType, left, right) =>
val buildSide = broadcastSideByHints(joinType, left, right)
Seq(joins.BroadcastHashJoinExec(
leftKeys, rightKeys, joinType, buildSide, condition, planLater(left), planLater(right)))
// broadcast hints were not specified, so need to infer it from size and configuration.
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if canBroadcastBySizes(joinType, left, right) =>
val buildSide = broadcastSideBySizes(joinType, left, right)
Seq(joins.BroadcastHashJoinExec(
leftKeys, rightKeys, joinType, buildSide, condition, planLater(left), planLater(right)))
// --- ShuffledHashJoin ---------------------------------------------------------------------
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if !conf.preferSortMergeJoin && canBuildRight(joinType) && canBuildLocalHashMap(right)
&& muchSmaller(right, left) ||
!RowOrdering.isOrderable(leftKeys) =>
Seq(joins.ShuffledHashJoinExec(
leftKeys, rightKeys, joinType, BuildRight, condition, planLater(left), planLater(right)))
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if !conf.preferSortMergeJoin && canBuildLeft(joinType) && canBuildLocalHashMap(left)
&& muchSmaller(left, right) ||
!RowOrdering.isOrderable(leftKeys) =>
Seq(joins.ShuffledHashJoinExec(
leftKeys, rightKeys, joinType, BuildLeft, condition, planLater(left), planLater(right)))
// --- SortMergeJoin ------------------------------------------------------------
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if RowOrdering.isOrderable(leftKeys) =>
joins.SortMergeJoinExec(
leftKeys, rightKeys, joinType, condition, planLater(left), planLater(right)) :: Nil
// --- Without joining keys ------------------------------------------------------------
// Pick BroadcastNestedLoopJoin if one side could be broadcasted
case j @ logical.Join(left, right, joinType, condition)
if canBroadcastByHints(joinType, left, right) =>
val buildSide = broadcastSideByHints(joinType, left, right)
joins.BroadcastNestedLoopJoinExec(
planLater(left), planLater(right), buildSide, joinType, condition) :: Nil
case j @ logical.Join(left, right, joinType, condition)
if canBroadcastBySizes(joinType, left, right) =>
val buildSide = broadcastSideBySizes(joinType, left, right)
joins.BroadcastNestedLoopJoinExec(
planLater(left), planLater(right), buildSide, joinType, condition) :: Nil
// Pick CartesianProduct for InnerJoin
case logical.Join(left, right, _: InnerLike, condition) =>
joins.CartesianProductExec(planLater(left), planLater(right), condition) :: Nil
case logical.Join(left, right, joinType, condition) =>
val buildSide = broadcastSide(
left.stats.hints.broadcast, right.stats.hints.broadcast, left, right)
// This join could be very slow or OOM
joins.BroadcastNestedLoopJoinExec(
planLater(left), planLater(right), buildSide, joinType, condition) :: Nil
// --- Cases where this strategy does not apply ---------------------------------------------
case _ => Nil
}
}
/**
* Used to plan streaming aggregation queries that are computed incrementally as part of a
* [[StreamingQuery]]. Currently this rule is injected into the planner
* on-demand, only when planning in a [[org.apache.spark.sql.execution.streaming.StreamExecution]]
*/
object StatefulAggregationStrategy extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case _ if !plan.isStreaming => Nil
case EventTimeWatermark(columnName, delay, child) =>
EventTimeWatermarkExec(columnName, delay, planLater(child)) :: Nil
case PhysicalAggregation(
namedGroupingExpressions, aggregateExpressions, rewrittenResultExpressions, child) =>
aggregate.AggUtils.planStreamingAggregation(
namedGroupingExpressions,
aggregateExpressions,
rewrittenResultExpressions,
planLater(child))
case _ => Nil
}
}
/**
* Used to plan the streaming deduplicate operator.
*/
object StreamingDeduplicationStrategy extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case Deduplicate(keys, child) if child.isStreaming =>
StreamingDeduplicateExec(keys, planLater(child)) :: Nil
case _ => Nil
}
}
object StreamingJoinStrategy extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = {
plan match {
case ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, condition, left, right)
if left.isStreaming && right.isStreaming =>
new StreamingSymmetricHashJoinExec(
leftKeys, rightKeys, joinType, condition, planLater(left), planLater(right)) :: Nil
case Join(left, right, _, _) if left.isStreaming && right.isStreaming =>
throw new AnalysisException(
"Stream stream joins without equality predicate is not supported", plan = Some(plan))
case _ => Nil
}
}
}
/**
* Used to plan the aggregate operator for expressions based on the AggregateFunction2 interface.
*/
object Aggregation extends Strategy {
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case PhysicalAggregation(
groupingExpressions, aggregateExpressions, resultExpressions, child) =>
val (functionsWithDistinct, functionsWithoutDistinct) =
aggregateExpressions.partition(_.isDistinct)
if (functionsWithDistinct.map(_.aggregateFunction.children).distinct.length > 1) {
// This is a sanity check. We should not reach here when we have multiple distinct
// column sets. Our MultipleDistinctRewriter should take care this case.
sys.error("You hit a query analyzer bug. Please report your query to " +
"Spark user mailing list.")
}
val aggregateOperator =
if (functionsWithDistinct.isEmpty) {
aggregate.AggUtils.planAggregateWithoutDistinct(
groupingExpressions,
aggregateExpressions,
resultExpressions,
planLater(child))
} else {
aggregate.AggUtils.planAggregateWithOneDistinct(
groupingExpressions,
functionsWithDistinct,
functionsWithoutDistinct,
resultExpressions,
planLater(child))
}
aggregateOperator
case _ => Nil
}
}
protected lazy val singleRowRdd = sparkContext.parallelize(Seq(InternalRow()), 1)
object InMemoryScans extends Strategy {
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case PhysicalOperation(projectList, filters, mem: InMemoryRelation) =>
pruneFilterProject(
projectList,
filters,
identity[Seq[Expression]], // All filters still need to be evaluated.
InMemoryTableScanExec(_, filters, mem)) :: Nil
case _ => Nil
}
}
/**
* This strategy is just for explaining `Dataset/DataFrame` created by `spark.readStream`.
* It won't affect the execution, because `StreamingRelation` will be replaced with
* `StreamingExecutionRelation` in `StreamingQueryManager` and `StreamingExecutionRelation` will
* be replaced with the real relation using the `Source` in `StreamExecution`.
*/
object StreamingRelationStrategy extends Strategy {
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case s: StreamingRelation =>
StreamingRelationExec(s.sourceName, s.output) :: Nil
case s: StreamingExecutionRelation =>
StreamingRelationExec(s.toString, s.output) :: Nil
case s: StreamingRelationV2 =>
StreamingRelationExec(s.sourceName, s.output) :: Nil
case _ => Nil
}
}
/**
* Strategy to convert [[FlatMapGroupsWithState]] logical operator to physical operator
* in streaming plans. Conversion for batch plans is handled by [[BasicOperators]].
*/
object FlatMapGroupsWithStateStrategy extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case FlatMapGroupsWithState(
func, keyDeser, valueDeser, groupAttr, dataAttr, outputAttr, stateEnc, outputMode, _,
timeout, child) =>
val execPlan = FlatMapGroupsWithStateExec(
func, keyDeser, valueDeser, groupAttr, dataAttr, outputAttr, None, stateEnc, outputMode,
timeout, batchTimestampMs = None, eventTimeWatermark = None, planLater(child))
execPlan :: Nil
case _ =>
Nil
}
}
// Can we automate these 'pass through' operations?
object BasicOperators extends Strategy {
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case d: DataWritingCommand => DataWritingCommandExec(d, planLater(d.query)) :: Nil
case r: RunnableCommand => ExecutedCommandExec(r) :: Nil
case MemoryPlan(sink, output) =>
val encoder = RowEncoder(sink.schema)
LocalTableScanExec(output, sink.allData.map(r => encoder.toRow(r).copy())) :: Nil
case MemoryPlanV2(sink, output) =>
val encoder = RowEncoder(StructType.fromAttributes(output))
LocalTableScanExec(output, sink.allData.map(r => encoder.toRow(r).copy())) :: Nil
case logical.Distinct(child) =>
throw new IllegalStateException(
"logical distinct operator should have been replaced by aggregate in the optimizer")
case logical.Intersect(left, right) =>
throw new IllegalStateException(
"logical intersect operator should have been replaced by semi-join in the optimizer")
case logical.Except(left, right) =>
throw new IllegalStateException(
"logical except operator should have been replaced by anti-join in the optimizer")
case logical.DeserializeToObject(deserializer, objAttr, child) =>
execution.DeserializeToObjectExec(deserializer, objAttr, planLater(child)) :: Nil
case logical.SerializeFromObject(serializer, child) =>
execution.SerializeFromObjectExec(serializer, planLater(child)) :: Nil
case logical.MapPartitions(f, objAttr, child) =>
execution.MapPartitionsExec(f, objAttr, planLater(child)) :: Nil
case logical.MapPartitionsInR(f, p, b, is, os, objAttr, child) =>
execution.MapPartitionsExec(
execution.r.MapPartitionsRWrapper(f, p, b, is, os), objAttr, planLater(child)) :: Nil
case logical.FlatMapGroupsInR(f, p, b, is, os, key, value, grouping, data, objAttr, child) =>
execution.FlatMapGroupsInRExec(f, p, b, is, os, key, value, grouping,
data, objAttr, planLater(child)) :: Nil
case logical.FlatMapGroupsInPandas(grouping, func, output, child) =>
execution.python.FlatMapGroupsInPandasExec(grouping, func, output, planLater(child)) :: Nil
case logical.MapElements(f, _, _, objAttr, child) =>
execution.MapElementsExec(f, objAttr, planLater(child)) :: Nil
case logical.AppendColumns(f, _, _, in, out, child) =>
execution.AppendColumnsExec(f, in, out, planLater(child)) :: Nil
case logical.AppendColumnsWithObject(f, childSer, newSer, child) =>
execution.AppendColumnsWithObjectExec(f, childSer, newSer, planLater(child)) :: Nil
case logical.MapGroups(f, key, value, grouping, data, objAttr, child) =>
execution.MapGroupsExec(f, key, value, grouping, data, objAttr, planLater(child)) :: Nil
case logical.FlatMapGroupsWithState(
f, key, value, grouping, data, output, _, _, _, timeout, child) =>
execution.MapGroupsExec(
f, key, value, grouping, data, output, timeout, planLater(child)) :: Nil
case logical.CoGroup(f, key, lObj, rObj, lGroup, rGroup, lAttr, rAttr, oAttr, left, right) =>
execution.CoGroupExec(
f, key, lObj, rObj, lGroup, rGroup, lAttr, rAttr, oAttr,
planLater(left), planLater(right)) :: Nil
case logical.Repartition(numPartitions, shuffle, child) =>
if (shuffle) {
ShuffleExchangeExec(RoundRobinPartitioning(numPartitions), planLater(child)) :: Nil
} else {
execution.CoalesceExec(numPartitions, planLater(child)) :: Nil
}
case logical.Sort(sortExprs, global, child) =>
execution.SortExec(sortExprs, global, planLater(child)) :: Nil
case logical.Project(projectList, child) =>
execution.ProjectExec(projectList, planLater(child)) :: Nil
case logical.Filter(condition, child) =>
execution.FilterExec(condition, planLater(child)) :: Nil
case f: logical.TypedFilter =>
execution.FilterExec(f.typedCondition(f.deserializer), planLater(f.child)) :: Nil
case e @ logical.Expand(_, _, child) =>
execution.ExpandExec(e.projections, e.output, planLater(child)) :: Nil
case logical.Window(windowExprs, partitionSpec, orderSpec, child) =>
execution.window.WindowExec(windowExprs, partitionSpec, orderSpec, planLater(child)) :: Nil
case logical.Sample(lb, ub, withReplacement, seed, child) =>
execution.SampleExec(lb, ub, withReplacement, seed, planLater(child)) :: Nil
case logical.LocalRelation(output, data, _) =>
LocalTableScanExec(output, data) :: Nil
case logical.LocalLimit(IntegerLiteral(limit), child) =>
execution.LocalLimitExec(limit, planLater(child)) :: Nil
case logical.GlobalLimit(IntegerLiteral(limit), child) =>
execution.GlobalLimitExec(limit, planLater(child)) :: Nil
case logical.Union(unionChildren) =>
execution.UnionExec(unionChildren.map(planLater)) :: Nil
case g @ logical.Generate(generator, _, outer, _, _, child) =>
execution.GenerateExec(
generator, g.requiredChildOutput, outer,
g.qualifiedGeneratorOutput, planLater(child)) :: Nil
case _: logical.OneRowRelation =>
execution.RDDScanExec(Nil, singleRowRdd, "OneRowRelation") :: Nil
case r: logical.Range =>
execution.RangeExec(r) :: Nil
case r: logical.RepartitionByExpression =>
exchange.ShuffleExchangeExec(r.partitioning, planLater(r.child)) :: Nil
case ExternalRDD(outputObjAttr, rdd) => ExternalRDDScanExec(outputObjAttr, rdd) :: Nil
case r: LogicalRDD =>
RDDScanExec(r.output, r.rdd, "ExistingRDD", r.outputPartitioning, r.outputOrdering) :: Nil
case h: ResolvedHint => planLater(h.child) :: Nil
case _ => Nil
}
}
}