Question
DataGrip 2023.3 Migration Rollback Safety: Automatic vs Manual Inverse Scripts
Lax Byte
0 reputation · 17 May 2025, 14:09 UTC
78.1K views1
Loading…
0 reputation · 17 May 2025, 14:09 UTC
The goal is to define the default behavior for rollback script generation in DataGrip’s migration tool when handling straightforward DDL changes such as adding a column or altering a data type. The decision must preserve the existing safety guarantee that only reversible changes are marked safe, while avoiding accidental data loss from automatically produced inverses that may not match the user’s intent.
Constraints include the version‑specific rollback safety checks introduced in DataGrip 2023.3, the need to keep the feature usable for teams that rely on explicit rollback scripts, and the requirement to maintain compatibility with older migration histories that lack automatic generation.
Should DataGrip automatically generate an inverse script for simple DDL operations, or should it require users to supply their own rollback? If automatic generation is adopted, what criteria should determine which changes are considered simple enough for safe auto‑generation?
DataGrip should automatically generate inverse (rollback) scripts for simple, reversible DDL changes, but require users to supply their own rollback when the migration involves data‑preserving steps, custom logic, or any ambiguous reversal.
datagrip
jetbrainsGenerate rollback script box.Does the migration include any data‑preserving or custom transformation (e.g., data migration, column split/merge, type change that could lose data)? If yes, a manual rollback script is safer; otherwise automatic generation is appropriate.
Use comments to ask for clarification. Post a solution as an answer.
5,940 reputation · 17 May 2025, 17:08 UTC
A critical distinction must be made between structural reversibility and data safety when configuring automatic inverse scripts. While DataGrip 2023.3 can structurally reverse an ADD COLUMN operation by generating a DROP COLUMN, this is a destructive action if the column has already been populated with production data.
To maintain safety guarantees, the tool should ideally flag any automatic inverse that utilizes a DROP or TRUNCATE command for manual review. For type alterations, verification is required to ensure the original precision or format is cached; otherwise, a structural reversal (e.g., INT back to VARCHAR) may succeed, but the original data state is lost.
DROP, DELETE).