What's new in Replicate May 2026?

This section describes the new and enhanced features in Replicate May 2026.

Customers should also review the Replicate release notes in Qlik Community for information about the following:

• Migration and upgrade
• End of life/support features
• Deprecated versions
• Resolved issues
• Known issues

New IBM DB2 for LUW target endpoint

This release adds IBM DB2 for LUW as a replication target, making it easy to replicate data from any supported source into Db2 for operational reporting and analytics

Using IBM DB2 for LUW as a target

New Cloudera Iceberg target endpoint

You can now replicate data from any supported source into Cloudera Iceberg. This release extends our replication targets to Iceberg tables in Cloudera, making it easier to land replicated datasets into an open table format for downstream analytics and lakehouse workloads.

Using Cloudera Iceberg as a target

New IBM IMS source endpoint

Replicate May 2026 introduces a new IBM IMS source endpoint that effectively replaces the legacy ARC (Attunity Replicate Connect) product. The new endpoint requires IBM IMS Connect to control access to IMS.

Key benefits:

• Minimal footprint and resource utilization on z/OS (primarily I/O and network; minimal CPU consumption)
• Modern security using mTLS for authentication and encryption
• High performance
• Minimal configuration required on z/OS
• Support for IMS catalog metadata
• Support for custom metadata preparation using an IBM toolset (for example, IBM IMS Explorer for Development)
• Uses the same IMS XXXXX Exit and LOGSTREAM facility as the legacy ARC IMS endpoint, meaning that both can run simultaneously during migration

Enhancements to MS-CDC based source endpoints

This version introduces several enhancements to MS-CDC based source endpoints (Google Cloud SQL for SQL Server, Microsoft Azure SQL (MS-CDC), and Microsoft SQL Server (MS-CDC)).

Enhanced support for replication of column subsets

In the past, customers who wanted to exclude specific columns from CDC could set a "Remove column" table transformation in Replicate.

However, Replicate would enable CDC on all the source table columns, instead of just the included columns. This would significantly impact performance and database resources. From this version, Replicate will use the @captured_column_list parameter to enable CDC on the included columns only. Alternatively, customers who do not want to define a transformation in Replicate can set the @captured_column_list parameter manually before starting the task.

Support for replicating tables with encrypted columns

In the past, tables with encrypted columns would be suspended. From this version, such tables will be written to the target without the encrypted columns.

Support for replicating tables with newly added columns

In the past, tables with newly added columns (ADD COLUMN DDL) would be suspended. Now such tables will be replicated to the target without the newly added column(s) and an assertion will be written to log.

SAP HANA source endpoint enhancements

This version introduces the following enhancements to the SAP HANA source endpoint:

CDC artifact cleanup

In Trigger-Based CDC mode, Replicate requires certain artifacts to be created in the source database (either manually by the DBA or automatically by Replicate). To preserve database disk space, you can now configure a deletion policy for artifacts that are no longer needed.

CDC artifacts

CDC continuity on dropped triggers

SAP application upgrades frequently drop foreign triggers from HANA databases. Removing these triggers causes Replicate to switch from CDC to full table reloads. For very large tables, these reloads can take over three days, resulting in data outages in the target system. Enabling the new Continue CDC on dropped triggers option will maintain CDC continuity when the required triggers are temporarily missing from the database.

CDC continuity on dropped triggers

MongoDB source endpoint enhancements

This version introduces the following enhancements to the MongoDB source endpoint:

• Support for reading a change stream (CDC) with events exceeding 16 MB.

• Support for STRICT (default), RELAXED and EXTENDED JSON modes. The modes can be selected using the new JSON mode drop-down in the endpoint's Advanced tab.

• Support for formatting numbers without scientific notation. This functionality is available via the following options in the endpoint's Advanced tab.

  • For Double data type, avoid scientific notation if the number of digits does not exceed: Select this option if you want to preserve the original Double data type notation when the number of Double data digits type does not exceed the specified value.

  • For Decimal128 data type, avoid scientific notation if the number of digits does not exceed: Select this option if you want to preserve the original Decimal128 data type notation when the number of Double data digits type does not exceed the specified value.

  For more information, see Setting advanced connection properties

Support for streaming records into Google Cloud BigQuery target endpoint

The ability to stream records into Google BigQuery has been added to the Google Cloud BigQuery target endpoint. Customers can now select either Batch loading or Streaming from the new Loading method drop-down in the General tab. When Streaming selected, records will be streamed into the database in batches, using the committed stream for Full Load and the default stream for CDC. For more information on these stream types, see Stream data using the Storage Write API. The Streaming method has several benefits over the Batch Loading method including lower costs, better performance, and once-only delivery. However, make sure you are aware of the limitations and considerations before selecting Streaming.

For a full list of benefits, see Introduction to the BigQuery Storage Write API.

Microsoft Fabric target endpoint enhancements

Support for loading records into a Mirrored Database according to timestamp

In the past, files were named using a sequence number. This caused various issues such as multiple files being named with the same number and overriding each other, data inconsistency after stopping and resuming a task, and others. The switch to sequence-based loading resolves these issues while simplifying the loading process.

Support for adding columns

The ALTER TABLE ADD column DDL is now supported. Note that the newly added column(s) must be nullable.

Data type enhancements

Microsoft SQL Server-based source endpoint - JSON support

All Microsoft SQL Server-based endpoints now support the JSON data type, which will be mapped to the NCLOB  with JSON subtype data type on Replicate.

Databricks Lakehouse (Delta) target endpoint - VARIANT support

Replicate’s STRING, WSTRING, CLOB and NCLOB data types will now be mapped to VARIANT on Databricks, if the target column has a JSON sub-type and Databricks Runtime version is 15.4 LTS or later.

Microsoft Fabric - Support for VARBINARY (MAX) and VARCHAR (MAX) data types

The mapping for the VARBINARY (MAX) and VARCHAR (MAX) data types will be as follows:

Certificate and test connection support in the server-level proxy settings

The server-level proxy settings now support:

• Specifying a certificate (in the new CA path field)

• Testing the connection using the new URL field and Test Connection button.

Default proxy server

Added support for creating a metadata file on table creation to the File, Amazon S3 and Google Cloud Storage target endpoints

In previous versions, when the Create metadata files in the target folder option was enabled, Replicate would generate a metadata file (DFM) for each data file. The metadata would contain information about the task, the table structure, and the associated data file. Now, customers can also select the new Create a metadata file on table creation option, which creates a metadata file during table creation. The new metadata file contains information about the task and the table structure only. This is useful when running a task in Metadata only mode or when the source table is empty - both of which will result in no data files being transferred to target. By providing details of the table structure, the new metadata file will help customers prepare their downstream applications accordingly. To align with the new option, the existing Create metadata files in the target folder option has been renamed to Create a metadata file for each data file.

Support for browsing for a local file

A new Browse button has been added to endpoint fields that, in the past, required you to specify a path to a file (for example, a certificate) on the Replicate Server machine. Best practice is to use the Browse button, which allows you to browse for the file locally. This eliminates the need to place files in certain paths on the Replicate Server machine. After the file is selected, its contents will be Base64 encoded in the endpoint settings, allowing you to delete the file from the local machine. Although not best practice, use of paths in parameters is still supported.

PostgreSQL: Support for TOAST storage

As this was certified for Replicate November 2025 after its initial release, some customers might have missed the post-release announcement that Replicate now supports PostgreSQL TOAST storage.

Support for using OAuth authentication with Kafka and Confluent Cloud target endpoints

The Kafka and Confluent Cloud target endpoints now support OAuth authentication.

Changes to Salesforce (Streaming CDC) and Salesforce (Incremental Load) source endpoints

Salesforce (Incremental Loading): Support for connecting to the proxy using HTTPS

You can now configure the endpoint's proxy server settings to connect using HTTPS with a Trusted CA certificate.

API version updates

When using the Salesforce (Streaming CDC) or Salesforce (Incremental Load) source endpoints, Replicate now requires version 64.0 of SOAP API and Bulk API.

When using the Salesforce (Streaming CDC) source endpoint, Replicate also requires version 64.0 of Streaming API.

Amazon S3 target - Assume Role

The Switch role after assuming SAML role option in the Amazon S3 target endpoint settings has been renamed to Assume Role.

Newly supported endpoint and driver versions

Newly supported endpoint versions

The following source and target endpoint versions are now supported:

• Oracle 23ai was renamed to Oracle AI Database 26ai

• Microsoft SQL Server 2025

• PostgreSQL 18

• Databricks 17.3 LTS

• DB2 LUW 12.1

• MariaDB 11.8

Newly supported driver versions

The following driver versions are now supported:

• MySQL: MySQL ODBC 9.6 Unicode driver

• Snowflake: Snowflake ODBC driver 3.15.0 (64-bit)

• Google BigQuery: Simba ODBC driver version 3.1.6.1026

• Databricks: Databricks ODBC Driver 2.11.0

  Warning noteTo align with Databricks, the driver for working with the Databricks Lakehouse (Delta) or Databricks (Cloud Storage) endpoints has been rebranded. To prevent tasks from failing, customer working with the Databricks Lakehouse (Delta) or Databricks (Cloud Storage) endpoints must upgrade their driver to Databricks ODBC Driver 2.11.0 or later. If you are installing the driver on Linux, make sure to edit the odbcinst.ini file according to the instructions.

Newly supported Replicate platform versions

• Red Hat Enterprise Linux 10.x (64-bit)

End-of-life endpoint versions

The following endpoint versions are now end of life.

Source endpoint versions

• MySQL 8.0 and 8.1
• IBM DB2 for iSeries 7.3

Target endpoint versions

• Databricks 12.2 LTS
• IBM DB2 for LUW 11.1 and 11.5