Stitch Documentation
has moved!

Please update your bookmarks to https://www.stitchdata.com/docs

If you're not automatically redirected after 5 seconds, click here.

Redshift Data Warehouse v2 Release Notes

Important! Version 1.0/1.1 Deprecation
Versions 1.0 and 1.1 of the Redshift data warehouse have been deprecated as of January 1, 2017. If you're unable to access your account, you may have been using one of these deprecated versions.

To continue using Stitch, you'll need to create a new account.

Welcome to version 2 of Stitch's Redshift data warehouse!

Version 2 Features

We’re excited to announce a number of new features that will vastly improve your Stitch experience. Here’s a look at what’s included in this release:

  • Bid adieu to views. In previous versions, data would first be replicated into a “staging” schema (_rjm) and then views for the new data would be created or updated. In v2, data will be stored directly in the tables in your integration schemas.
  • No more staging schema. Because of this new replication process, the _rjm schema has been deprecated. This, coupled with storing data directly in integration schema tables, will drastically reduce issues with view creation and the time spent troubleshooting errors.
  • Increase querying efficiency. Want to speed up your queries? In this version, we added support for applying encodings, sort, and distribution keys.
  • What are these _sdc columns? These columns replace the staging _rjm schema from previous versions. Any column prepended with _sdc is used by Stitch to replicate your data.

    Additionally, in previous versions of Stitch, when tables were de-nested into subtablescolumns used as parts of composite keys - prepended with _rjm - would be added. These columns will continue to be added to tables, but will now be prepended with _sdc instead. For example: _rjm_source_key_id will be _sdc_source_key_id in version 2.

    Note that if any of your columns are prepended with _rjm or _sdc, they will not be replicated to your data warehouse. Stitch will reserve any columns of this form for internal use.

Version 1 Recap

While we’re at it, here’s a recap of what was released in v1:

  • Genuine multiple data type splits. If there is more than one data type in a single column, Stitch will split the data into multiple columns: one for each data type.
  • Automatic VARCHAR widening. Up to 65k characters, VARCHAR data types will automatically be widened to the smallest possible amount without truncation. This means if a data point with 65 characters comes in, Stitch will automatically widen the column to 128 to accommodate it.
  • Hashed column names are a thing of the past. Previously, Stitch would hash column names that only differed by case - consider Salesforce’s infamous use of id and Id - to prevent collisions. Now, data type names will be used instead.

    For example: an integration sends over two columns. One is named updated_at and contains DATE data types; the other is Updated_At and contains TIMESTAMPs. Stitch will add the data types to the column names, naming them updated_at__date and updated_at__timestamp, respectively.

Migrating Your Account

Switch your Redshift data warehouse over to v2 before January 1, 2017 by following the instructions in this article: Migrating to Redshift v2.

 

RELATED

Was this article helpful?
0 out of 0 found this helpful

Comments

Questions or suggestions? If something in our documentation is unclear, let us know in the comments!