The data types Stitch supports for replication fall into two categories:

  • Common, which are data types supported for all integrations
  • Integration-specific, which are data types supported for specific integrations and integration versions, where applicable. Note: Common data types also apply to integrations that support integration-specific data types.

Note: If a data type isn’t present in either the Common or Integration-specific tables, it means that Stitch doesn’t currently support replication for that data type. Replicating columns with unsupported data types may lead to issues with replication.

As of June 19, 2018, the data types in the table below are supported for all integrations.

BIGINT 1 BIT 2 CHAR 3 DATE 0
DECIMAL 1 DOUBLE 2 FLOAT 3 INTEGER 0
LONGVARCHAR 1 LONGNVARCHAR 2 NCHAR 3 NVARCHAR 0
NUMERIC 1 REAL 2 SMALLINT 3 TIME 0
TIMESTAMP 1 TINYINT 2 VARCHAR 3

PostgreSQL supported data types

PostgreSQL integrations support the data types in the table below, in addition to the common data types.

In the table are the data types and how they'll be stored in your destination, depending on the version of the integration being used:

  • marks supported data types. Following the icon is how Stitch will store the data type in the destination. For example: ARRAY data types will be stored as strings.
  • marks unsupported data types. In Stitch, columns with unsupported data types will have Unavailable in the Type column of the Tables to Replicate tab.
PostgreSQL
version 1.0 (latest) 		PostgreSQL
version 15-10-2015
COMMON DATA TYPES Supported Supported
ARRAY STRING STRING
CIDR STRING Unsupported
CITEXT STRING Unsupported
HSTORE Nested maps STRING
INET STRING Unsupported
IP STRING Unsupported
JSON STRING STRING
JSONB STRING STRING
MACADDR STRING Unsupported
MONEY STRING Unsupported
UUID STRING STRING

Columns with mixed data types

Occasionally Stitch will encounter a column that contains more than one data type. As Stitch requires that there be only one data type per column to properly type and store your data, columns containing multiple data types may be “split” to ensure all values are correctly typed.

For example: an order_confirmed column is synced and typed as BOOLEAN. In a subsequent sync, Stitch detects VARCHAR values in this column.

As a result, Stitch will create an additional column to accommodate the VARCHAR values. The new column name will be the original name appended with the data type: order_confirmed__string

How columns are named depends on the type of destination being used to warehouse data. Refer to the Mixed Data Types guide for more info.

Replicating unsupported data types

Columns containing data types that aren’t supported may prevent an entire table from replicating.

If you determine a non-replicating column contains an unsupported data type, you’ll need to de-select it to allow the table to successfully replicate.

‘Not Supported’ status for columns

If a column you want to sync has a Status of Not Supported, the root cause may be an unsupported data type or insufficient user permissions.

Related Troubleshooting
Questions? Feedback?

Did this article help? If you have questions or feedback, feel free to submit a pull request with your suggestions, open an issue on GitHub, or reach out to us.

Tags: replication