v0.4.0

This release solves a lot of issues and updates things to where Redshift and BigQuery are today, see #22 for more information. A big thanks to @gh-mlfowler for this release.

v0.3.2

This is a bugfix release that adds support for tables in schemas other than public (see #7, thanks to @iserko for the excellent patch), and tries to fix a problem with connection timeouts for very long running unloads (see #9).

v0.3.1

This release fixes a bug in v0.3.0 that disabled compression by default, when the documentation said that it was enabled by default.

v0.3.0

This version adds two new command line arguments:

• --[no-]compression makes it possible to skip compression to work around BigQuery's 4 GiB compressed files limit
• --steps makes it possible to run only some parts of a transfer, for example to resume a transfer when BigShift accidentally got killed during the transfer step, or to run the load step against multiple BigQuery datasets.

v0.2.0

This release includes many improvements and updates to the documentation, including:

• You can now set the maximum bad records to allow when loading the BigQuery table: --max-bad-records
• Specifying the BigQuery table ID is now optional, it will default to the same name as the Redshift table
• The transfer uses Redshift's unload manifest to determine which files to transfer, instead of relying on globs
• The unloaded and transferred files will be removed from S3 and Cloud Storage after the BigQuery table has been loaded
• The data will be compressed to minimize transfer costs
• The S3 and Cloud Storage locations specified are now required to be empty or the transfer will fail
• Logging has been improved, for example the number of files and their sizes are logged before the transfer starts

v0.1.1

This is the first (working) release of BigShift.

v0.1.0

This version has been yanked from RubyGems because of an error in the release.