ReferenceFlow: compare SQL Server reference data and generate a guarded sync script
Lookup, reference and configuration tables have to match between development, test and production. ReferenceFlow compares them from inside your JetBrains IDE, using the SQL Server data sources you already have, and turns the differences you select into a T-SQL script you review before anything runs.
Submitted to JetBrains Marketplace on 2026-10-23 and in review there.
- Validated with
- SQL Server 2022
- IDEs
- DataGrip 2025.1 runtime-tested; API checks passed for DataGrip 2026.2 and Rider 2026.2
- Size limit
- Up to 10,000 rows and 128 columns per table
- Requires
- Database Tools and SQL, SELECT on the compared tables, VIEW DEFINITION on the database
- Price
- Paid plugin on JetBrains Marketplace
What ReferenceFlow for SQL Server does
Compare by primary key
See inserts, updates, target-only rows and unchanged counts. Text keys are matched with the target's SQL Server collation.
Inspect what differs
Changed columns are shown with the primary key kept visible. Switch to all columns, or press F4 to see the full original value, with an optional escaped-character view.
Pick each change
Filter the row differences and choose exactly which operations go into the script. Hidden selections are still counted.
Deletes are opt-in
Deletion and explicit identity insertion are off by default. Target-only rows stay protected until you turn deletion on.
Guarded script
The script checks the server, database and table metadata, verifies the original selected rows and applies the changes in one transaction. Detected conflicts stop it.
You run it
Copy the script, save it as a file or open it in the target's SQL console. Opening the console does not execute it.
Saved profiles
Comparison setups keep connection identifiers and table names, without passwords, captured rows or generated SQL.
Good to know
- It compares table data, not schemas. It is not a schema migration or cross-database deployment tool.
- Source and target need compatible column definitions and matching primary keys. Oversized or unsupported input is rejected instead of giving a partial comparison.
- Views, tables with triggers, temporal and ledger tables, row-level security, masked or encrypted columns, and XML, spatial, sql_variant and user-defined types are outside this release.
- Azure SQL and SQL Server 2025 have not been runtime-validated. Foreign keys stay enforced, and dependent tables are not synchronized automatically.
Privacy
ReferenceFlow runs inside your JetBrains IDE and connects through the data sources you configured there.
Comparing reads both tables and changes neither database. The plugin never applies a synchronization script by itself.
Saved comparison profiles keep connection identifiers and table names only: no passwords, captured rows or generated SQL.
Purchases and licenses are handled by JetBrains Marketplace under JetBrains' own terms and privacy policy.
FAQ
Does ReferenceFlow change my database?
Comparing doesn't: it only reads both tables. Changes happen only when you run the generated script yourself, and the script stops if it finds conflicts.
Can it delete rows in the target?
Only if you turn deletion on. It is off by default, and target-only rows stay protected until then.
Does it compare schemas?
No. It compares and synchronizes table data. Columns and primary keys have to match already.
How big can the tables be?
Up to 10,000 rows and 128 columns per table, within cell, snapshot and script-size limits.
Related tools
Updated 2026-10-23