You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@atlas.apache.org by "Umesh Padashetty (Jira)" <ji...@apache.org> on 2020/12/10 20:10:00 UTC
[jira] [Updated] (ATLAS-4072) spark_column_lineage missing for
insert into select * queries run via spark-shell
[ https://issues.apache.org/jira/browse/ATLAS-4072?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Umesh Padashetty updated ATLAS-4072:
------------------------------------
Attachment: Screenshot 2020-12-11 at 1.29.39 AM.png
> spark_column_lineage missing for insert into select * queries run via spark-shell
> ---------------------------------------------------------------------------------
>
> Key: ATLAS-4072
> URL: https://issues.apache.org/jira/browse/ATLAS-4072
> Project: Atlas
> Issue Type: Bug
> Components: atlas-core
> Reporter: Umesh Padashetty
> Priority: Major
> Attachments: Screenshot 2020-12-11 at 1.29.39 AM.png, Screenshot 2020-12-11 at 1.34.24 AM.png
>
>
> From the spark-shell, ran the below queries
> * spark.sql("create table umesh(name string)");
> * spark.sql("create table umesh_insert(name string)");
> * spark.sql("insert into umesh_insert select * from umesh");
> There is a spark_process created between umesh and umesh_insert tables, but the spark_column_lineage is missing between the umesh.name and umesh_insert.name columns
> !Screenshot 2020-12-11 at 1.29.39 AM.png|width=438,height=435!
> To cross verify the behavior, I ran similar hive queries via beeline and found out that along with hive_process being created between umesh_hive and umesh_hive_insert tables, hive_column_lineage is created between umesh_hive.name and umesh_hive_insert.name columns.
> Queries run via beeline
> * create table umesh_hive(name string);
> * create table umesh_hive_insert(name string);
> * insert into umesh_hive_insert select * from umesh_hive;
> !Screenshot 2020-12-11 at 1.34.24 AM.png|width=441,height=548!
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)