You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "Ravindra Pesala (JIRA)" <ji...@apache.org> on 2018/03/26 04:17:00 UTC

[jira] [Created] (CARBONDATA-2278) Save the datamaps to system folder of warehouse

Ravindra Pesala created CARBONDATA-2278:
-------------------------------------------

             Summary: Save the datamaps to system folder of warehouse
                 Key: CARBONDATA-2278
                 URL: https://issues.apache.org/jira/browse/CARBONDATA-2278
             Project: CarbonData
          Issue Type: New Feature
            Reporter: Ravindra Pesala


*Problem*
Currently, datamap schema is saved inside the main table schema itself. This approach cannot satisfy if a datamap belongs to more than one table. For suppose if we need to create a datamap joining 2 tables then we cannot keep the datamap schema under any one table.
And also accessing the datamaps required to read the main table schema every time, it is not well optimized. And if we need to create multiple datamaps for a table then all datamaps need to store under the schema of that table so the size of main table schema grows and impacts the performance.

*Solution*
Make the datamap schema independent of main table schema. And store the schema under {{_system}}folder location. This location is configurable by using carbon property {{carbon.system.folder.location}} , by default, it stores under the store location.
Created datamap schema in JSON format for better readability. And has the interfaces to store it in database.
Made {{on table <tablename>}} for datamap DDL as optional , so now user can create/drop or show datamaps without {{on table}} option.



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