Skip to content

Releases: tonyabasy/flink-sql-bootstrap

Release 1.0.1

18 Jun 16:29
@tonyabasy tonyabasy
v1.0.1
ad1b8f0

Choose a tag to compare

View all tags
Release 1.0.1 Latest
Latest

Changed

  • Domain-driven resource profile renaming — Resource profiles renamed from cup-size naming to operator-responsibility naming (small → stateless, normal → stateful, large → join_heavy, xlarge → sink), with resource values tuned accordingly. Not backward-compatible with legacy resource configuration JSON

Added

  • Documentation site — Built a documentation site with bilingual architecture, including blogs, user guide, and community pages

Compatibility

Flink Local YARN-App YARN-Session K8s-Session K8s-App
2.2.0 ✅ ✅ ✅ ✅ ✅
2.1.1 ✅ ✅ ✅ ✅ ✅
2.0.2 ✅ ✅ ✅ ✅ ✅
1.20.4 ✅ ✅ ✅ ✅ ✅

20/20 PASS · 0 failures

Assets 3
Loading

Flink SQL Bootstrap v1.0.0

08 Jun 05:52
@tonyabasy tonyabasy
v1.0.0
98ed1fa

Choose a tag to compare

View all tags
Flink SQL Bootstrap v1.0.0

Added

  • Multi-Statement SQL Script execution engine — Supports mixing CREATE TABLE, SET, INSERT, CALL, and other statements in a single .sql file. Statements are automatically split, validated, and orchestrated in semantic order, with DML statements deferred to the compilation phase for unified submission.
  • Catalog snapshot pre-registration — Pre-register tables, views, and UDFs via a JSON file so the Catalog is fully ready at job startup, eliminating the need for DDL in the SQL script. Supports multi-protocol loading: classpath:, file://, http(s)://, and hdfs://.
  • Fine-grained operator-level resource injection — Inject parallelism, CPU, heap memory, managed memory, off-heap memory, external resources, and chain strategy at the Transformation DAG level, matched by operator UID or name. Operators with identical resource configs are automatically grouped into the same SlotSharingGroup.
  • SQL syntax validation (--validate) — Validate SQL syntax locally without submitting to a Flink cluster. Parse errors include exact line and column numbers for rapid iteration.
  • SQL compilation and execution plan output (--compile) — Parse, validate, and compile SQL, outputting the InternalPlan JSON execution plan without actually submitting the job, useful for preview and debugging.
  • Resource template auto-generation (--init-resource) — Automatically extract the Transformation DAG structure from the current SQL script and generate a per-operator resource configuration JSON template. Users can modify values and inject them directly.
  • Flink 1.20.x / 2.x dual-version compatibility — Bypass SPI compatibility checks via ApplicationOperationExecutor, and fix the URI→URL type conversion ArrayStoreException via UriSafeSessionContext, enabling normal job startup in Application Mode.
  • Deterministic operator UID generation — Force-enable TABLE_EXEC_UID_GENERATION = ALWAYS to ensure every Transformation has a stable UID, serving as the exact match key in the resource configuration JSON.
  • Formatted SQL result printing — Compatible with Flink 1.20.x/2.x TableauStyle result table rendering, with exponential backoff via RowDataIterator for polling results.
  • Exception hierarchy — Define SqlValidateException, SqlCompileException, and SqlParsePosException; SQL parse errors carry source line and column position information.
  • Experimental API marker — Introduce the @Experimental annotation to mark incubating APIs such as the DAG printer.
  • DAG topology visualization (experimental) — ASCII-art rendering of the Transformation DAG, supporting typical topologies such as dual-source Join, Union, and multi-way aggregation in the console.

Build & CI

  • Maven-based build with Shade Plugin configured for fat JAR packaging; main class is SqlEntryPoint.
  • Spotless configured for code formatting and automatic Apache 2.0 License Header injection.
  • GitHub Actions CI configured with Java 11 / 17 / 21 matrix builds.
  • 45 unit tests covering core executor, resource injection, entry-point security, and resource spec signing logic.

Documentation

  • Complete Chinese and English READMEs, including Quick Start, CLI options, and configuration examples.

  • Capability boundary documentation (CAPABILITIES.md).

  • Domain glossary (CONTEXT.md).

  • SQL examples (example-word-count.sql, example-word-count-advanced.sql).

  • Catalog snapshot example (example-catalog.json).

  • Operator resource configuration example (example-resource.json).

  • Sample UDF JARs (example-udf-reverse.jar, example-udf-substring.jar).

  • AI Agent collaboration guidelines (docs/agents/).

  • Flink multi-version compatibility test suite (scripts/flink-compat-test/) — automated testing across Flink 1.20 ~ 2.2 in Local, YARN, and Kubernetes deployment modes. Generates HTML compatibility reports (docs/flink-compat-test-<version>.html) with pass/fail matrix and error classification.

  • All 5 deployment modes (Local, YARN-App, YARN-Session, K8s-Session, K8s-App) tested and passing.

  • Full compatibility report: docs/flink-compat-test-1.0.0.html

Compatibility

Flink Local YARN-App YARN-Session K8s-Session K8s-App
2.2.0 ✅ ✅ ✅ ✅ ✅
2.1.1 ✅ ✅ ✅ ✅ ✅
2.0.2 ✅ ✅ ✅ ✅ ✅
1.20.4 ✅ ✅ ✅ ✅ ✅

20/20 PASS · 0 failures

Loading