SQL type detection
Detect query, DML, DDL, SHOW and EXPLAIN statement types.
15+ datasource dialects supported
MySQL
Hive
Doris
StarRocks
ClickHouse
Oracle
PostgreSQL
SparkTrino
DM
OceanBaseOceanBase Oracle
SQL Server
Greenplum
TiDBCapabilities
One engine for the full SQL parsing pipeline
From dialect syntax to business semantics — unified contract, unified result, ready to use.
Table & DDL parsing
Resolve real tables through JOINs, CTEs and subqueries, then parse fields, keys, partitions and buckets.
Table lineage
Trace data flow from source tables to targets in INSERT and CTAS statements.
Column lineage
Map source columns to targets across expressions, functions and CASE clauses.
Function parsing
Extract and deduplicate regular, aggregate and window function names across multiple statements.
Parameter extraction & prepared SQL
Extract dynamic SELECT parameters and produce JDBC placeholder SQL with ordered bindings.
Architecture
Two-layer unified extension architecture
The syntax layer extends dialect Lexer/Parser/AST on Druid; the business layer extends parsing, lineage and rewriting via the ISqlParser contract, aligned by DbType and AST.
- 1Feed SQL + DbType
- 2Syntax layer parses into Druid AST
- 3Business layer unified parsing
- 4Table / lineage / function / parameter results
Documentation
Version / i18n coexistence
Switch version (1.0 / Next) and language (简体中文 / English) from the navbar.
Ready to parse your SQL?
Get started in minutes with multi-dialect parsing, lineage tracing and prepared SQL.