Hi Aniket, I think it will be better if we can consider delete and update compaction feature also in the high level design list.
Thanks, Sujith On Nov 23, 2016 4:29 AM, "Aniket Adnaik (JIRA)" <j...@apache.org> wrote: > Aniket Adnaik created CARBONDATA-440: > ---------------------------------------- > > Summary: Provide Update/Delete functionality support in > CarbonData > Key: CARBONDATA-440 > URL: https://issues.apache.org/jira/browse/CARBONDATA-440 > Project: CarbonData > Issue Type: New Feature > Components: core, data-query, file-format, spark-integration, sql > Affects Versions: 0.1.1-incubating, 0.1.0-incubating, 0.2.0-incubating > Reporter: Aniket Adnaik > Fix For: 0.3.0-incubating > > > Currently, CarbonData does not support modification of existing rows in > the table. This is a major limitation for many real world desirable use > cases in data warehousing, such as slow changing dimension tables, data > correction of fact tables or data cleanup, etc. Many users want to be able > to update and delete rows from the CarbonData table. > > Following are some high level design goals to support this functionality, > 1. Provide a standard SQL interface to perform Update and Delete > operations. > 2. Perform Update and Delete operations on CarbonData table without having > to rewrite the entire CarbonData block (file) by making use of differential > files (a.k.a delta files). > 3. After Update or Delete operation, CarbonData readers should skip > deleted records and read updated records seamlessly without having to > modify user applications. > > > > -- > This message was sent by Atlassian JIRA > (v6.3.4#6332) >
