You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Chris O'Hara (JIRA)" <ji...@apache.org> on 2018/08/06 23:20:00 UTC

[jira] [Updated] (SPARK-25037) plan.transformAllExpressions doesn't transform expressions in subquery plans

     [ https://issues.apache.org/jira/browse/SPARK-25037?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chris O'Hara updated SPARK-25037:
---------------------------------
    Description: 
Given the following LogicalPlan, containing a SubqueryAlias and SubqueryExpression:
{code:java}
scala> val plan = spark.sql("SELECT 1 bar FROM (SELECT 1 foo) WHERE foo IN (SELECT 1 foo)").queryExecution.logical
plan: org.apache.spark.sql.catalyst.plans.logical.LogicalPlan =
'Project [1 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [1 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [1 AS foo#28]
         +- OneRowRelation
{code}
The following transformation should replace all instances of lit(1) with lit(2):
{code:java}
scala> plan.transformAllExpressions { case l @ Literal(1, _) => l.copy(value = 2) }
res0: plan.type =
'Project [2 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [1 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [2 AS foo#28]
         +- OneRowRelation
{code}
Instead, the nested SubqueryExpression plan is not transformed.

The expected output is: 
{code:java}
'Project [2 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [2 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [2 AS foo#28]
         +- OneRowRelation
{code}
 

 

  was:
Given the following LogicalPlan, containing a SubqueryAlias and SubqueryExpression:

 
{code:java}
scala> val plan = spark.sql("SELECT 1 bar FROM (SELECT 1 foo) WHERE foo IN (SELECT 1 foo)").queryExecution.logical
plan: org.apache.spark.sql.catalyst.plans.logical.LogicalPlan =
'Project [1 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [1 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [1 AS foo#28]
         +- OneRowRelation
{code}
The following transformation should replace all instances of lit(1) with lit(2):

 
{code:java}
scala> plan.transformAllExpressions { case l @ Literal(1, _) => l.copy(value = 2) }
res0: plan.type =
'Project [2 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [1 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [2 AS foo#28]
         +- OneRowRelation
{code}
 

Instead, the nested SubqueryExpression plan is not transformed.

The expected output is:

 
{code:java}
'Project [2 AS bar#29]
+- 'Filter 'foo IN (list#31 [])
   :  +- Project [2 AS foo#30]
   :     +- OneRowRelation
   +- SubqueryAlias __auto_generated_subquery_name
      +- Project [2 AS foo#28]
         +- OneRowRelation
{code}
 

 

 

 


> plan.transformAllExpressions doesn't transform expressions in subquery plans
> ----------------------------------------------------------------------------
>
>                 Key: SPARK-25037
>                 URL: https://issues.apache.org/jira/browse/SPARK-25037
>             Project: Spark
>          Issue Type: Bug
>          Components: SQL
>    Affects Versions: 2.3.1
>            Reporter: Chris O'Hara
>            Priority: Minor
>
> Given the following LogicalPlan, containing a SubqueryAlias and SubqueryExpression:
> {code:java}
> scala> val plan = spark.sql("SELECT 1 bar FROM (SELECT 1 foo) WHERE foo IN (SELECT 1 foo)").queryExecution.logical
> plan: org.apache.spark.sql.catalyst.plans.logical.LogicalPlan =
> 'Project [1 AS bar#29]
> +- 'Filter 'foo IN (list#31 [])
>    :  +- Project [1 AS foo#30]
>    :     +- OneRowRelation
>    +- SubqueryAlias __auto_generated_subquery_name
>       +- Project [1 AS foo#28]
>          +- OneRowRelation
> {code}
> The following transformation should replace all instances of lit(1) with lit(2):
> {code:java}
> scala> plan.transformAllExpressions { case l @ Literal(1, _) => l.copy(value = 2) }
> res0: plan.type =
> 'Project [2 AS bar#29]
> +- 'Filter 'foo IN (list#31 [])
>    :  +- Project [1 AS foo#30]
>    :     +- OneRowRelation
>    +- SubqueryAlias __auto_generated_subquery_name
>       +- Project [2 AS foo#28]
>          +- OneRowRelation
> {code}
> Instead, the nested SubqueryExpression plan is not transformed.
> The expected output is: 
> {code:java}
> 'Project [2 AS bar#29]
> +- 'Filter 'foo IN (list#31 [])
>    :  +- Project [2 AS foo#30]
>    :     +- OneRowRelation
>    +- SubqueryAlias __auto_generated_subquery_name
>       +- Project [2 AS foo#28]
>          +- OneRowRelation
> {code}
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org