Apache Flink SQL Parser Support in GSP

Explore the Apache Flink SQL Parser product page

Apache Flink SQL is the SQL layer of the Apache Flink stream-processing framework, used to define streaming and batch pipelines with standard SQL plus streaming extensions such as CREATE TABLE ... WITH (...) connector definitions, watermarks, and windowing table-valued functions.

General SQL Parser (GSP) parses Flink SQL with a dedicated grammar, so Flink jobs can be analyzed offline: full AST access, syntax validation, formatting, and column-level data lineage across streaming pipelines. This makes GSP suitable for building lineage and impact analysis over Flink-based data platforms.

See the Flink SQL Keyword Compatibility Reference for keyword-as-identifier support.

Syntax Coverage Corpus

GSP tests Flink SQL syntax against 39 documented constructs grouped into 11 categories, each cross-referenced to the documentation that defines it (22 distinct source pages).

Parse coverage

Measured against GSP Java 4.1.11 on 2026-08-10 (UTC) (artifact sha256 71895b6e5ab0). Each construct below carries the complete set of examples from the Flink SQL documentation; a construct counts as fully covered only when every one of its examples parses. Percentages are pass rates over this documented-example corpus — not a percentage of the entire SQL dialect.

	Constructs fully parsed	Individual statements parsed
All constructs	38 of 39 (97%)	127 of 129 (98%)
SELECT (queries)	12 of 13 (92%)	42 of 44 (95%)
DML (INSERT / UPDATE / DELETE / MERGE)	2 of 2 (100%)	6 of 6 (100%)
Stored procedures	1 of 1 (100%)	5 of 5 (100%)

Statements are split using the parser's own statement splitter, so multi-statement examples and procedure bodies are counted correctly. The SELECT, DML and stored-procedure rows classify each statement by its parsed statement type: a construct is counted in a row when its examples contain at least one statement of that class, and is fully parsed when all of them parse. Stored procedures covers CREATE PROCEDURE / FUNCTION / PACKAGE / TRIGGER definitions and anonymous blocks.

Machine-readable: dialect data (JSON) · all-dialect matrix · coverage matrix overview

All 39 constructs

Construct	Category	Parses	Specified in
ALTER DATABASE	ALTER	all 1	Flink docs
ALTER FUNCTION	ALTER	all 2	Flink docs
ALTER TABLE	ALTER	all 6	Flink docs
ALTER VIEW	ALTER	all 2	Flink docs
CREATE CATALOG	CREATE	all 2	Flink docs
CREATE DATABASE	CREATE	all 2	Flink docs
CREATE FUNCTION	CREATE	all 5	Flink docs
CREATE TABLE	CREATE	all 6	Flink docs
CREATE TABLE AS SELECT	CREATE	all 2	Flink docs
CREATE TABLE LIKE	CREATE	all 2	Flink docs
CREATE VIEW	CREATE	all 2	Flink docs
DESCRIBE	DESCRIBE	all 5	Flink docs
DROP CATALOG	DROP	all 1	Flink docs
DROP DATABASE	DROP	all 2	Flink docs
DROP FUNCTION	DROP	all 2	Flink docs
DROP TABLE	DROP	all 2	Flink docs
DROP VIEW	DROP	all 2	Flink docs
EXPLAIN	EXPLAIN	all 3	Flink docs
EXECUTE STATEMENT SET	INSERT	all 3	Flink docs
INSERT	INSERT	all 5	Flink docs
RESET	RESET	all 2	Flink docs
SELECT	SELECT	all 6	Flink docs
SELECT Group Aggregation	SELECT	all 5	Flink docs
SELECT JOIN	SELECT	all 6	Flink docs
SELECT Lateral Table Join	SELECT	all 2	Flink docs
SELECT LIMIT	SELECT	all 1	Flink docs
SELECT MATCH_RECOGNIZE	SELECT	0 of 2	Flink docs
SELECT ORDER BY	SELECT	all 1	Flink docs
SELECT Over Aggregation	SELECT	all 3	Flink docs
SELECT Set Operations	SELECT	all 6	Flink docs
SELECT Top-N	SELECT	all 1	Flink docs
SELECT Window Aggregation	SELECT	all 5	Flink docs
SELECT Windowing TVF (TUMBLE, HOP, CUMULATE, SESSION)	SELECT	all 6	Flink docs
SELECT WITH (Common Table Expression)	SELECT	all 1	Flink docs
SET	SET	all 2	Flink docs
SHOW	SHOW	all 6	Flink docs
SHOW CREATE, SHOW FUNCTIONS	SHOW	all 6	Flink docs
SHOW TABLES	SHOW	all 6	Flink docs
USE	USE	all 5	Flink docs