ddlike

ddlike is a conceptual diff-based language and tooling designed to describe incremental changes to structured data. It is used to express patches that can be applied to documents such as JSON, YAML, or tabular representations, enabling reproducible updates, data migrations, and synchronization workflows. In ddlike, a patch consists of a sequence of operations that describe how the target data should be transformed.

The core operations typically include add, remove, replace, move, and copy. Each operation targets a data path, which is a slash-separated notation identifying a location within the document. Some proposals for ddlike extend the path syntax with support for arrays, wildcards, or conditional predicates, allowing patches to apply conditionally or to multiple elements in a single step. Patches may also include values for insertion or replacement and can be structured to be invertible, aiding undo or rollback processes.

Semantics of patch application are designed to be deterministic: applying the same patch to the same document yields the same result, with well-defined behavior for conflicts and missing targets. Conflict resolution rules can be built into the patch engine, including last-writer-wins, user prompts, or merge strategies that consider patch provenance.

ddlike is not an official standard but exists primarily in academic discussions and experimental tooling as a way to compare diff-based data transformations with line-oriented or JSON Patch approaches. It has been explored as a means to support data versioning, incremental migration, and cross-system synchronization. Related concepts include JSON Patch and general data-diff methodologies.