Homogeneous replication
The following section describes how Replicate handles replication from Google Cloud SQL for PostgreSQL to the following targets: PostgreSQL, Microsoft Azure Database for PostgreSQL, Google Cloud SQL for PostgreSQL, and Google Cloud AlloyDB for PostgreSQL.
In homogeneous replication, source column and table collations will be replicated to the target as described in Column and table collation.
In homogeneous replication, the source data first passes through the Qlik Replicate data type and is therefore subject to any limitations of that type.
For information on Replicate data types and their limitations (where relevant), see Replicate data types.
For information on which Replicate data types the source data passes through when replicating from Google Cloud SQL for PostgreSQL, see the Google Cloud SQL for PostgreSQL to Qlik Replicate data types mapping table described earlier.
Limitations and considerations
Homogeneous replication is not supported with the following Replicate configurations or operations:
- Apply Changes only tasks
- Stopping and resuming a task
- Starting a task from a timestamp
- When the If target table already exists in the task settings' Full Load Settings tab is set to Do nothing
Data type considerations and exceptions
When replicating from Google Cloud SQL for PostgreSQL to a PostgreSQL-based target, source and target data types will be identical, with the following exceptions:
| Google Cloud SQL for PostgreSQL source | PostgreSQL-based target |
|---|---|
|
ENUM |
STRING |
|
COMPOSITE |
STRING |
|
NUMERIC |
NUMERIC (28,6) |
|
NUMERIC (p,s) A numeric array with precision and scale. |
NUMERIC A numeric array without precision and scale. |
| TINTERVAL |
Information note
As PostgreSQL have discontinued support for the TINTERVAL data type, tables with this data type will not be created on the target if the target database version does not support it. |
| INT4MULTIRANGE | STRING (255) |
| INT8MULTIRANGE | STRING (255) |
| NUMMULTIRANGE | STRING (255) |
| TSMULTIRANGE | STRING (255) |
See also the note about JSON data type replication in Supported data types.
Column and table collation
When replicating from Google Cloud SQL for PostgreSQL to a PostgreSQL-based target, column and table collations will be replicated to the target.
To support collation replication, the DBA must ensure that the collations defined for the Google Cloud SQL for PostgreSQL database are the same as those defined for the PostgreSQL-based target database.
Non-nullable columns and Primary/Unique index names
Non-nullable columns and Primary/Unique Index names are preserved during homogeneous replication.
Replication of Partitioned Tables
In order to replicate partitioned tables to a PostgreSQL-based target, you need to do the following:
- Manually create the parent and child tables in the target database.
- Define a separate task to replicate those tables.
- Set the Truncate before loading option in the Full Load Settings.
- If the Support partitioned tables in CDC option in the Advanced tab is not selected, select the child tables only (not the partition table).