spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Davies Liu (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (SPARK-14757) Incorrect behavior of Join operation in Spqrk SQL JOIN : "false" in the left table is joined to "null" on the right table
Date Sat, 30 Apr 2016 05:29:12 GMT

     [ https://issues.apache.org/jira/browse/SPARK-14757?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Davies Liu updated SPARK-14757:
-------------------------------
    Assignee: Reynold Xin

> Incorrect behavior of Join operation in Spqrk SQL JOIN : "false" in the left table is
joined to "null" on the right table
> -------------------------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-14757
>                 URL: https://issues.apache.org/jira/browse/SPARK-14757
>             Project: Spark
>          Issue Type: Bug
>          Components: SQL
>    Affects Versions: 1.6.0
>            Reporter: Hong Huang
>            Assignee: Reynold Xin
>             Fix For: 1.6.2, 2.0.0
>
>
> Content of table a:
> |outgoing_0|
> |     false     |
> |      true     |
> |      null      |
> a has only one field: outgoing_0 
> Content of table b:
> |outgoing_1|
> |     false      |
> |      true      |
> |      null       |
> b has only one filed: outgoing_1
> After running this query:
> {code}
> select * from a FULL JOIN b ON ( outgoing_0<=>outgoing_1)
> {code}
> I got the following result:
> |outgoing_0|outgoing_1|
> |      true      |      true      |
> |     false      |     false     |
> |     false      |      null      |
> |      null       |      null      |
> The row with "false" as outgoing_0 and "null" as outgoing_1 is unexpected. The operator
<=> should match null with null. 
> While left "false" is matched with right "null", it is also strange to find that the
"false" on the right table does not match with "null" on the left table (no row with "null"
as outgoing_0 and "false" as outgoing_1)
> You can easily reproduce this bug by pasting the following code fragment:
> {code}
> case class A( outgoing_0: Option[Boolean] )
> case class B( outgoing_1: Option[Boolean] )
> {code}
> {code}
> val a = sc.parallelize( Seq(
>       A( Some( false ) ),
>       A( Some( true ) ),
>       A( None )
>     ) ).toDF()
>     a.show
>     val b = sc.parallelize( Seq(
>       B( Some( false ) ),
>       B( Some( true ) ),
>       B( None )
>     ) ).toDF()
>     b.show
>     a.registerTempTable( "a" )
>     b.registerTempTable( "b" )
>     sqlContext.sql( "select * from a FULL JOIN b ON ( outgoing_0<=>outgoing_1)"
).show()
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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


Mime
View raw message