Product and documentation updates about the PostgreSQL integration from the team at Stitch.
Stay up-to-date by subscribing to our RSS feed.
PostgreSQL (v2) update: Google CloudSQL PostgreSQL now available!
A new version (v2) of our Google CloudSQL PostgreSQL integration is now available!
Along with all of the new features in v2 of the PostgreSQL integration (which this integration is based on), we’ve also added support for:
- SSL connections
- Log-based Incremental Replication
Learn more in our Google CloudSQL PostgreSQL integration documentation.
PostgreSQL (v2) update: Leaving beta!
A new version (v) of our integration has left beta and is now generally available!
This version (v) of Stitch’s integration optimizes replication by utilizing Avro schemas to write and validate data, thereby reducing the amount of time spent on data extraction and preparation. Compared to previous versions of the integration, this version boasts increased performance and overall reduced replication time.
Notable improvements and changes in this version also include:
- Expanded data type support. This version supports additional data types. Refer to the data types documentation for more info.
- Improved handling of
JSON,JSONB, andHSTOREdata types. In previous versions, these data types were treated as strings. This version will send them to your destination as JSON objects, which may result in de-nesting. - Improved handling of schema changes in tables using Log-based Incremental Replication. Adding and removing columns in these tables will no longer cause extraction errors. Note: This limitation still applies to version 1.
Note: The following features aren’t fully supported, but are being worked on:
- Arrays of
DECIMAL,NUMERIC, andTIMESTAMP(TZ)data types - **Support for other flavors of **. We’re in the process of testing this version of our integration with other flavors of , including Heroku, Google CloudSQL, Amazon Aurora, and Amazon RDS.
To get a look at how this version compares to the previous version of , refer to the version comparison documentation.
PostgreSQL (HP) (v2) integration: Removed support for Composite, Range and OID types
We’ve removed support for Composite Types, Range Types and Object Identifier Types from our integration.
PostgreSQL (v2) update: Key-based Incremental Replication now available!
The v version of our integration, which is currently in open beta, now supports Key-based Incremental Replication. All of Stitch’s supported Replication Methods are now available for this version - go try it out!
New feature: Integration changelogs
Introducing: Integration changelogs!
Check out the history of our integrations and stay in the loop on updates with dedicated changelogs for these integrations:
We’ll add changelogs for other popular integrations and destinations in the weeks to come. If there’s a specific integration you’d like to see us work on, let us know by creating an issue in the Stitch Docs GitHub repo.
PostgreSQL integration: New version (v2) now in beta
A new version (v) of our integration is now in beta!
This version (v) of Stitch’s integration optimizes replication by utilizing Avro schemas to write and validate data, thereby reducing the amount of time spent on data extraction and preparation. Compared to previous versions of the integration, this version boasts increased performance and overall reduced replication time.
Notable improvements and changes in this version also include:
- New column (field) naming rules. Avro has specific rules that dictate how columns can be named. As a result, column names will be canonicalized to adhere to Avro rules and persisted to your destination using the Avro-friendly name. Refer to the Column name transformations section in the docs for more info.
- Expanded data type support. This version supports additional data types. Refer to the data types documentation for more info.
- Improved handling of
JSON,JSONB, andHSTOREdata types. In previous versions, these data types were treated as strings. This version will send them to your destination as JSON objects, which may result in de-nesting.
Note: The following features aren’t currently supported, but will be before the integration leaves beta:
- Key-based Incremental Replication
ARRAYdata type
To get a look at how this version compares to the previous version of , refer to the version comparison documentation.
PostgreSQL (v1) integrations: Include partitioned tables in discovery
Previously, partitioned tables weren’t being detected during Discovery for (v) integrations. This has been corrected.
PostgreSQL (v1) integrations: Correctly use SSL connections
We’ve fixed an issue with (v) integrations where SSL connections weren’t being used, even if the SSL option was checked in Stitch. If checked, Stitch will now correctly use and enforce SSL when connecting to the database.
PostgreSQL (v1) integrations: Correctly display BYTEA data as unsupported
Previously, Stitch would display columns typed as BYTEA as available for selection, despite this data type currently being unsupported for (v) integrations. With this fix, these columns will now correctly display as UNSUPPORTED in Stitch.
PostgreSQL (v15-10-2015) sunset
The v version of our integration has been sunset. As of today, this version of the integration has been removed from Stitch and will no longer function.
Migrate to the latest version of (v1) today to continue replicating data.
PostgreSQL (v1) integrations: Correctly type BIGINT[] data for Log-based Incremental Replication
Previously, Log-based Incremental Replication was incorrectly typing data types of INT8[] as STRING during data type conversion. This has been corrected.
PostgreSQL (v15-10-2015) deprecation
The v version of our integration has been deprecated. As of today, this version of the integration is no longer formally supported.
While connections using v will continue to run, this version will be sunset in the future. Migrate to the latest version of (v1) today to prevent possible disruptions.
PostgreSQL (v1) integrations: BIGINT array support
(v) integrations now support arrays of BIGINTs.
PostgreSQL (v1) integrations: Fix selected-by-default behavior
We’ve fixed an issue with (v) integrations where columns marked as selected-by-default will now be replicated unless explicitly deselected.
PostgreSQL (v1) integrations: Primary Key discovery
Previously, Primary Keys weren’t detected during Discovery if other indices or unique constraints existed on the column. This has been corrected.
PostgreSQL (v1) integrations: Include schema name in destination table names
We’ve added a new setting to (v) integrations:
When checked, Stitch will include schema names from the source database in the destination table name when the table is created. For example: <source_schema_name>__<table_name>
Stitch loads all selected replicated tables to a single schema, preserving only the table name. If two tables canonicalize to the same name - even if they’re in different source databases or schemas - name collision errors can arise. Checking this setting can prevent these issues.
Note: This setting can not be changed after the integration is saved. Additionally, this setting may create table names that exceed your destination’s limits.
PostgreSQL (v1) integrations: Expanded replication slot support
Stitch now supports Log-based Incremental Replication across multiple databases in a single cluster. Previously, Stitch supported one replication slot per integration and required it to be named stitch. Now, Stitch will look for the replication slot you define before falling back to stitch.
PostgreSQL (v1) integrations: PG 10 now supported for Logical Replication
Stitch now supports 10 when using Log-based Incremental Replication. Previously, Stitch supported 9.4.x - 9.9.x.
PostgreSQL integration: New version (v1)
A new version (v) of our integration is now available!
We’ve worked hard to ensure this new integration is the best way to extract data from to your Stitch destination. The new integration, based on the Singer standard, includes many new features such as:
- Logical replication support, which allows you to replicate data from database binary logs, including hard deletes, without the use of a Replication Key
- User-facing Extraction Logs
- Table and field selection
- Enhanced scheduling options
- Run and stop Extraction on demand functionality
- Availability via the Stitch Connect API
- Enhanced schema validation
Get started today by creating a new integration or learn more in the updated documentation.
Google Cloud SQL PostgreSQL Support
We’re excited to report that Stitch supports the recently announced PostgreSQL flavor of the the Google CloudSQL service. It’s available as both an input integration AND a destination - a dual threat!
Reset Replication Keys for database integrations
Need to completely re-replicate all incrementally replicated tables in a database integration? Now you can(!) via the Integration Settings page on any MySQL, PostgreSQL, Microsoft SQL Server, or MongoDB integration. Take care though - the re-replicated rows count towards your monthly caps.
PostgreSQL (v15-10-2015) integrations: ARRAY and JSONB data type support
All fields of type JSONB and ARRAY may now be tracked from the Integrations page. These fields will be converted to strings in Amazon Redshift.