This section describes how you interact through Trifacta® SaaS with your Redshift data warehouse.
- Redshift is a scalable data warehouse solution available through Amazon. For more information, see https://aws.amazon.com/redshift/.
- No schema validation is performed as part of writing results to Redshift.
- Credentials and permissions are not validated when you are modifying the destination for a publishing job.
- For Redshift, no validation is performed to determine if the target is a view and is therefore not a supported target.
Uses of Redshift
The Trifacta platform can use Redshift for the following tasks:
- Create datasets by reading from Redshift tables.
Write to Redshift tables with your job results.
Ad-hoc publication of data to Redshift.
Before You Begin Using Redshift
Read Access: Your Redshift administrator must configure read permissions. Your administrator should provide a database for upload to your Redshift datastore.
Write Access: You can write and publish jobs results to Redshift.
SSL is required.
Storing Data in Redshift
Your Redshift administrator should provide database access for storing datasets. Users should know where shared data is located and where personal data can be saved without interfering with or confusing other users.
NOTE: Trifacta SaaS does not modify source data in Redshift. Datasets sourced from Redshift are read without modification from their source locations.
Reading from Redshift
You can create a Trifacta dataset from a table or view stored in Redshift.
NOTE: The Redshift cluster must be in the same region as the default S3 bucket.
NOTE: If a Redshift connection has an invalid iamRoleArn, you can browse, import datasets, and open the data in the Transformer page. However, any jobs executed using this connection fail. If the iamRoleArn is invalid, the only samples that you can generate are Quick Random samples; other sampling jobs fail.
For more information, see Redshift Browser.
Writing to Redshift
NOTE: You cannot publish to a Redshift database that is empty. The database must contain at least one table.
You can write back data to Redshift using one of the following methods:
- Job results can be written directly to Redshift as part of the normal job execution. Create a new publishing action to write to Redshift. See Run Job Page.
As needed, you can publish results to Redshift for previously executed jobs.
NOTE: You cannot re-publish results to Redshift if the original job published to Redshift. However, if the dataset was transformed but publication to Redshift failed, you can publish from the Publishing dialog.
NOTE: To publish to Redshift, the source results must be in Avro or JSON format.
- For more information on how data is converted to Redshift, see Redshift Data Type Conversions.
Data Validation issues:
- No validation is performed for the connection and any required permissions during job execution. So, you can be permitted to launch your job even if you do not have sufficient connectivity or permissions to access the data. The corresponding publish job fails at runtime.
- Prior to publication, no validation is performed on whether a target is a table or a view, so the job that was launched fails at runtime.
This page has no comments.