[ 
https://issues.apache.org/jira/browse/CARBONDATA-224?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15472826#comment-15472826
 ] 

ASF GitHub Bot commented on CARBONDATA-224:
-------------------------------------------

Github user gvramana commented on a diff in the pull request:

    https://github.com/apache/incubator-carbondata/pull/139#discussion_r77947424
  
    --- Diff: 
integration/spark/src/main/scala/org/apache/spark/sql/CarbonSqlParser.scala ---
    @@ -864,8 +864,8 @@ class CarbonSqlParser()
        /**
         * detects whether double or decimal column is part of 
dictionary_exclude
         */
    -  def isDoubleDecimalColDictionaryExclude(columnDataType: String): Boolean 
= {
    -    val dataTypes = Array("double", "decimal")
    +  def isNumberInColDictionaryExclude(columnDataType: String): Boolean = {
    +     val dataTypes = Array("double", "decimal", "int")
    --- End diff --
    
    Add a reverse check that only string or (varchar-check) is only supported. 
As other datatypes like bigint etc will have problem in above check


> Fixed data mismatch issue in case of Dictionary Exclude column for Numeric 
> data type
> ------------------------------------------------------------------------------------
>
>                 Key: CARBONDATA-224
>                 URL: https://issues.apache.org/jira/browse/CARBONDATA-224
>             Project: CarbonData
>          Issue Type: Bug
>            Reporter: kumar vishal
>            Assignee: kumar vishal
>
> Problem: In case of greater than query on dictionary exclude column of for 
> numeric data type
> This is because data is sorted based on string because of this if data is 
> 1,10,2,3 , data will be sorted like 1,10,2,3 but if we search greater than 3 
> then while applying min max will return false as last value is 3
> Solution:we need to sort based on actual data type for this we should have 
> chain comparator based on data type while loading the data, currently 
> disabling DictionaryExclude column for numeric data type and will throw 
> exception. Will raise jira issue to for sorting the based on actual data type



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

Reply via email to