Re: update bug with carbondata1.1.0 and spark1.6.0

Posted by Liang Chen-2 on Jun 20, 2017; 2:25am
URL: http://apache-carbondata-dev-mailing-list-archive.168.s1.nabble.com/update-bug-with-carbondata1-1-0-and-spark1-6-0-tp15578p15587.html

Hi

Just i copied your code and tested it at my local machine , the result as
below. Because there are no any "india" value in name column, so don't do
any update.  Please check it again.
BTW, suggest you do test at branch-1.1,  because the community is testing
branch-1.1 for preparing 1.1.1 patch release, the "update and delete" is
key improvement in 1.1.1 release.

+-----+
|  _c0|
+-----+
|80000|


+-----+
|  _c0|
+-----+
|80000|
+-----+

Regards
Liang

2017-06-20 9:38 GMT+08:00 yangwei <[hidden email]>:

> Hi,
>   My test case is as follows:
>
> import  cc.implicits._
> var kk:Seq[(String,String,String,Int)] = Nil
> for(i<-1 to 80000){
>    kk = kk ++Seq((i+"", "name","city", i))
> }
> cc.sc.makeRDD(kk).toDF("id","name","city","age").registerTempTable("t1")
> cc.sql("DROP TABLE IF EXISTS testtable")
> cc.sql("CREATE TABLE IF NOT EXISTS testtable(id string, name string, city
> string, age Int) STORED BY 'carbondata'")
> cc.sql("insert INTO TABLE testtable select * from t1")
>   cc.sql("""
>          UPDATE testtable SET (testtable.name) = ('india') WHERE
> testtable.id = '10'
>          """).show()
> cc.sql("SELECT count(distinct id) FROM testtable").show()
> |-------|
> |---c0-|
> |80000|
>
> cc.sql("""
>         UPDATE testtable SET (testtable.name) = ('india') WHERE
> testtable.id
> = '10'
>         """).show()
>  cc.sql("SELECT count(distinct id) FROM testtable").show()
> |-------|
> |---c0-|
> |79998|
> Only one record is updated, but the total number of records is missing by
> two
>
>
>
> --
> View this message in context: http://apache-carbondata-dev-
> mailing-list-archive.1130556.n5.nabble.com/update-bug-with-
> carbondata1-1-0-and-spark1-6-0-tp15578.html
> Sent from the Apache CarbonData Dev Mailing List archive mailing list
> archive at Nabble.com.
>