When will the MERGE SQL query be supported in Microsoft Fabric Warehouse?

[FabricTrailLear]

Hello,

I am currently working with Microsoft Fabric Warehouse and attempting to use the MERGE SQL statement to perform upserts (insert/update operations). However, it seems that the MERGE statement is currently not supported in Fabric's warehouse SQL engine.

Could someone from the Microsoft team or the community provide clarification on:

• Whether MERGE support is planned for Microsoft Fabric Warehouse?
• If so, is there an estimated timeline for when this feature might be rolled out?
• Are there any recommended workarounds in the meantime for achieving similar functionality?

This feature is quite essential for many data integration and ETL/ELT scenarios, and I’m sure many users would benefit from knowing the roadmap or alternatives.

Thanks in advance for your help!

[rosha_rosha]

MERGE support is planned as part of ongoing SQL capabilities improvements in Fabric Warehouse, but:

• No official ETA has been provided publicly by Microsoft.
• Based on trends, it is expected sometime in 2025, possibly in H2 2025, as part of broader T-SQL language parity updates.

You can follow updates on:

• Microsoft Fabric Ideas Forum
• Official Microsoft Learn release wave plans

You can follow updates on:

• Microsoft Fabric Ideas Forum
• Official Microsoft Learn release wave plans

Workarounds for MERGE in Fabric Warehouse

Until MERGE is available, you can replicate its behavior with manual UPSERT logic using a combination of UPDATE and INSERT with NOT EXISTS.

-- 1. UPDATE existing records

UPDATE target

SET target.column1 = source.column1,

    target.column2 = source.column2

FROM target_table AS target

JOIN staging_table AS source

    ON target.primary_key = source.primary_key;

-- 2. INSERT new records

INSERT INTO target_table (primary_key, column1, column2)

SELECT source.primary_key, source.column1, source.column2

FROM staging_table AS source

LEFT JOIN target_table AS target

    ON target.primary_key = source.primary_key

WHERE target.primary_key IS NULL;

Tip: If your tables are large, consider indexing your join keys and/or breaking up operations into batches for performance.

Alternative: ELT with Dataflows Gen2 + Notebooks

If you're orchestrating pipelines in Fabric:

• Use Dataflows Gen2 for staging and deduplication
• Combine with Lakehouse Notebooks (PySpark) for flexible logic
• Write the final curated data into your Warehouse