[ 
https://issues.apache.org/jira/browse/ATLAS-2532?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Ashutosh Mestry updated ATLAS-2532:
-----------------------------------
    Attachment: ATLAS-2532-dsl-max-float.patch

> Advanced Search DSL: Improve Support for Numeric Data Types at Thresholds
> -------------------------------------------------------------------------
>
>                 Key: ATLAS-2532
>                 URL: https://issues.apache.org/jira/browse/ATLAS-2532
>             Project: Atlas
>          Issue Type: Bug
>          Components:  atlas-core
>    Affects Versions: trunk
>            Reporter: Ashutosh Mestry
>            Assignee: Ashutosh Mestry
>            Priority: Major
>         Attachments: ATLAS-2532-dsl-max-float.patch
>
>
> Steps to duplicate the problem:
>  * Added this property to _fs_path_
>                 _{_
>                     _"name": "retention",_
>                     _"typeName": "float",_
>                     _"cardinality": "SINGLE",_
>                     _"isIndexable": false,_
>                     _"isOptional": true,_
>                     _"isUnique": false_
>                 _}_
>  * From Atlas WEB UI, add _hdfs_path_ entity. Set the _retention_ property to 
> 3.4028235e+38.
>  * From Atlas WEB UI, Adanced Search use: hdfs_path where retention > 
> 3.4028235e+38
> Actual behavior: Error _‘__Gremlin script execution failed: 
> g.V().has('__typeName', 'hdfs_path').has('hdfs_path.retention', 
> gt(3.4028235e+38)).dedup().limit(25).toList()’_ is displayed.
> Expected behavior: Results should be displayed.
> Root cause:
>  * Since the value is on threshold of max float, Gremlin converts this to 
> _BigDecimal_ and then does the comparision.
>  
> Fix:
>  * Using _‘f’_ to suppress conversion to _BigDecimal_ should address the 
> problem.
>  * Using _g.V().has('__typeName', 'hdfs_path').has('hdfs_path.retention', 
> lte(3.4028235e+38f)).dedup().limit(25).toList()_ addresses the problem.
>  



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

Reply via email to