We currently haven't pulled very much of the ACID utilities into the ORC
project, although we plan to. Currently you would need to write the
extended schema and put them into directories with the correct names. You
need the Hive metastore for assigning transaction ids (eg, assigning your
writer transaction 1234 so that it knows to write into
delta_001234_001234/bucket_000000) and to commit the transaction after you
are done writing.

On Wed, Aug 24, 2016 at 6:46 AM, Stephen Carman <[email protected]>
wrote:

> Hi everyone,
>
> I'm trying to create a basic ACID Orc file from the orc library. I am able
> to write basic Orc files fine, but I am struggling to see how I would add
> the ACID features to the files. Do I just alter the schema to include the
> ACID stats? Also, the documentation talks about the Hive metastore in some
> places, is having a Hive metastore necessary for the ACID transactions to
> work? A simple example of how to write and then update an ACID ORC file
> would be amazing.
>
> Thanks!
> --
> Stephen Carman, M.S.
> Research Engineer, Planalytics
> (M) 267.240.0363
> [email protected]
> www.planalytics.com
>

Reply via email to