Skip to content

[SPARK-33850][SQL] EXPLAIN FORMATTED doesn't show the plan for subqueries if AQE is enabled #30855

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 6 commits into from
Closed

[SPARK-33850][SQL] EXPLAIN FORMATTED doesn't show the plan for subqueries if AQE is enabled #30855

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
165cd91
Fix an issue that EXPLAIN FORMATTED doesn't show subquery when APE en…
sarutak Dec 18, 2020
35f127a
Add test.
sarutak Dec 19, 2020
2ee8d19
Modify test.
sarutak Dec 19, 2020
6cbe6a6
Add JIRA ID to the test.
sarutak Dec 19, 2020
577dd8f
Change the variable name "adp" to "a"
sarutak Dec 19, 2020
13fb225
Modify explain-aqe.sql.out to comply with the change.
sarutak Dec 19, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/ExplainUtils.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -218,6 +218,8 @@ object ExplainUtils extends AdaptiveSparkPlanHelper {
plan: => QueryPlan[_],
subqueries: ArrayBuffer[(SparkPlan, Expression, BaseSubqueryExec)]): Unit = {
plan.foreach {
case a: AdaptiveSparkPlanExec =>
getSubqueries(a.executedPlan, subqueries)
case p: SparkPlan =>
p.expressions.foreach (_.collect {
case e: PlanExpression[_] =>
Expand Down
263 changes: 263 additions & 0 deletions sql/core/src/test/resources/sql-tests/results/explain-aqe.sql.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -407,6 +407,101 @@ Condition : (((isnotnull(key#x) AND isnotnull(val#x)) AND (key#x = Subquery subq
Output [2]: [key#x, val#x]
Arguments: isFinalPlan=false

===== Subqueries =====

Subquery:1 Hosting operator id = 2 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (10)
+- HashAggregate (9)
+- Exchange (8)
+- HashAggregate (7)
+- Project (6)
+- Filter (5)
+- Scan parquet default.explain_temp2 (4)


(4) Scan parquet default.explain_temp2
Output [2]: [key#x, val#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp2]
PushedFilters: [IsNotNull(key), IsNotNull(val), EqualTo(val,2)]
ReadSchema: struct<key:int,val:int>

(5) Filter
Input [2]: [key#x, val#x]
Condition : (((isnotnull(key#x) AND isnotnull(val#x)) AND (key#x = Subquery subquery#x, [id=#x])) AND (val#x = 2))

(6) Project
Output [1]: [key#x]
Input [2]: [key#x, val#x]

(7) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_max(key#x)]
Aggregate Attributes [1]: [max#x]
Results [1]: [max#x]

(8) Exchange
Input [1]: [max#x]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(9) HashAggregate
Input [1]: [max#x]
Keys: []
Functions [1]: [max(key#x)]
Aggregate Attributes [1]: [max(key#x)#x]
Results [1]: [max(key#x)#x AS max(key)#x]

(10) AdaptiveSparkPlan
Output [1]: [max(key)#x]
Arguments: isFinalPlan=false

Subquery:2 Hosting operator id = 5 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (17)
+- HashAggregate (16)
+- Exchange (15)
+- HashAggregate (14)
+- Project (13)
+- Filter (12)
+- Scan parquet default.explain_temp3 (11)


(11) Scan parquet default.explain_temp3
Output [2]: [key#x, val#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp3]
PushedFilters: [IsNotNull(val), GreaterThan(val,0)]
ReadSchema: struct<key:int,val:int>

(12) Filter
Input [2]: [key#x, val#x]
Condition : (isnotnull(val#x) AND (val#x > 0))

(13) Project
Output [1]: [key#x]
Input [2]: [key#x, val#x]

(14) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_max(key#x)]
Aggregate Attributes [1]: [max#x]
Results [1]: [max#x]

(15) Exchange
Input [1]: [max#x]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(16) HashAggregate
Input [1]: [max#x]
Keys: []
Functions [1]: [max(key#x)]
Aggregate Attributes [1]: [max(key#x)#x]
Results [1]: [max(key#x)#x AS max(key)#x]

(17) AdaptiveSparkPlan
Output [1]: [max(key)#x]
Arguments: isFinalPlan=false

-- !query
EXPLAIN FORMATTED
Expand Down Expand Up @@ -442,6 +537,101 @@ Condition : ((key#x = Subquery subquery#x, [id=#x]) OR (cast(key#x as double) =
Output [2]: [key#x, val#x]
Arguments: isFinalPlan=false

===== Subqueries =====

Subquery:1 Hosting operator id = 2 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (10)
+- HashAggregate (9)
+- Exchange (8)
+- HashAggregate (7)
+- Project (6)
+- Filter (5)
+- Scan parquet default.explain_temp2 (4)


(4) Scan parquet default.explain_temp2
Output [2]: [key#x, val#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp2]
PushedFilters: [IsNotNull(val), GreaterThan(val,0)]
ReadSchema: struct<key:int,val:int>

(5) Filter
Input [2]: [key#x, val#x]
Condition : (isnotnull(val#x) AND (val#x > 0))

(6) Project
Output [1]: [key#x]
Input [2]: [key#x, val#x]

(7) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_max(key#x)]
Aggregate Attributes [1]: [max#x]
Results [1]: [max#x]

(8) Exchange
Input [1]: [max#x]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(9) HashAggregate
Input [1]: [max#x]
Keys: []
Functions [1]: [max(key#x)]
Aggregate Attributes [1]: [max(key#x)#x]
Results [1]: [max(key#x)#x AS max(key)#x]

(10) AdaptiveSparkPlan
Output [1]: [max(key)#x]
Arguments: isFinalPlan=false

Subquery:2 Hosting operator id = 2 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (17)
+- HashAggregate (16)
+- Exchange (15)
+- HashAggregate (14)
+- Project (13)
+- Filter (12)
+- Scan parquet default.explain_temp3 (11)


(11) Scan parquet default.explain_temp3
Output [2]: [key#x, val#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp3]
PushedFilters: [IsNotNull(val), GreaterThan(val,0)]
ReadSchema: struct<key:int,val:int>

(12) Filter
Input [2]: [key#x, val#x]
Condition : (isnotnull(val#x) AND (val#x > 0))

(13) Project
Output [1]: [key#x]
Input [2]: [key#x, val#x]

(14) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_avg(cast(key#x as bigint))]
Aggregate Attributes [2]: [sum#x, count#xL]
Results [2]: [sum#x, count#xL]

(15) Exchange
Input [2]: [sum#x, count#xL]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(16) HashAggregate
Input [2]: [sum#x, count#xL]
Keys: []
Functions [1]: [avg(cast(key#x as bigint))]
Aggregate Attributes [1]: [avg(cast(key#x as bigint))#x]
Results [1]: [avg(cast(key#x as bigint))#x AS avg(key)#x]

(17) AdaptiveSparkPlan
Output [1]: [avg(key)#x]
Arguments: isFinalPlan=false

-- !query
EXPLAIN FORMATTED
Expand Down Expand Up @@ -470,6 +660,79 @@ Input: []
Output [1]: [(scalarsubquery() + scalarsubquery())#x]
Arguments: isFinalPlan=false

===== Subqueries =====

Subquery:1 Hosting operator id = 2 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (8)
+- HashAggregate (7)
+- Exchange (6)
+- HashAggregate (5)
+- Scan parquet default.explain_temp1 (4)


(4) Scan parquet default.explain_temp1
Output [1]: [key#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp1]
ReadSchema: struct<key:int>

(5) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_avg(cast(key#x as bigint))]
Aggregate Attributes [2]: [sum#x, count#xL]
Results [2]: [sum#x, count#xL]

(6) Exchange
Input [2]: [sum#x, count#xL]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(7) HashAggregate
Input [2]: [sum#x, count#xL]
Keys: []
Functions [1]: [avg(cast(key#x as bigint))]
Aggregate Attributes [1]: [avg(cast(key#x as bigint))#x]
Results [1]: [avg(cast(key#x as bigint))#x AS avg(key)#x]

(8) AdaptiveSparkPlan
Output [1]: [avg(key)#x]
Arguments: isFinalPlan=false

Subquery:2 Hosting operator id = 2 Hosting Expression = Subquery subquery#x, [id=#x]
AdaptiveSparkPlan (13)
+- HashAggregate (12)
+- Exchange (11)
+- HashAggregate (10)
+- Scan parquet default.explain_temp1 (9)


(9) Scan parquet default.explain_temp1
Output [1]: [key#x]
Batched: true
Location [not included in comparison]/{warehouse_dir}/explain_temp1]
ReadSchema: struct<key:int>

(10) HashAggregate
Input [1]: [key#x]
Keys: []
Functions [1]: [partial_avg(cast(key#x as bigint))]
Aggregate Attributes [2]: [sum#x, count#xL]
Results [2]: [sum#x, count#xL]

(11) Exchange
Input [2]: [sum#x, count#xL]
Arguments: SinglePartition, ENSURE_REQUIREMENTS, [id=#x]

(12) HashAggregate
Input [2]: [sum#x, count#xL]
Keys: []
Functions [1]: [avg(cast(key#x as bigint))]
Aggregate Attributes [1]: [avg(cast(key#x as bigint))#x]
Results [1]: [avg(cast(key#x as bigint))#x AS avg(key)#x]

(13) AdaptiveSparkPlan
Output [1]: [avg(key)#x]
Arguments: isFinalPlan=false

-- !query
EXPLAIN FORMATTED
Expand Down
22 changes: 22 additions & 0 deletions sql/core/src/test/scala/org/apache/spark/sql/ExplainSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -277,6 +277,28 @@ class ExplainSuite extends ExplainSuiteHelper with DisableAdaptiveExecutionSuite
}
}

test("SPARK-33850: explain formatted - check presence of subquery in case of AQE") {
withTable("df1") {
withSQLConf(SQLConf.ADAPTIVE_EXECUTION_ENABLED.key -> "true") {
withTable("df1") {
spark.range(1, 100)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think its better to use temporary views in tests where possible.

.write
.format("parquet")
.mode("overwrite")
.saveAsTable("df1")
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: df1 -> df (this is a nit comment, so I think we don't need a follow-up pr to fix it)


val sqlText = "EXPLAIN FORMATTED SELECT (SELECT min(id) FROM df1) as v"
val expected_pattern1 =
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: we don't the number in the name, I think. expected_pattern1 -> expected_pattern

"Subquery:1 Hosting operator id = 2 Hosting Expression = Subquery subquery#x"

withNormalizedExplain(sqlText) { normalizedOutput =>
assert(expected_pattern1.r.findAllMatchIn(normalizedOutput).length == 1)
}
}
}
}
}

test("Support ExplainMode in Dataset.explain") {
val df1 = Seq((1, 2), (2, 3)).toDF("k", "v1")
val df2 = Seq((2, 3), (1, 1)).toDF("k", "v2")
Expand Down