sqlglot-rust

A SQL parser, optimizer, and transpiler library written in Rust, inspired by Python's sqlglot.

Features

Parse SQL strings into a structured AST
Generate SQL from AST nodes
Transpile between 30 SQL dialects
C/C++ FFI — shared & static libraries with a C header for integration from C, C++, or any language with C ABI support
CLI — command-line interface for transpiling, parsing, and formatting SQL
Expression Builder API — fluent builders for programmatic SQL construction
Typed function expressions — 72+ functions across 8 categories with dialect-specific generation
LIMIT / TOP / FETCH FIRST transpilation across dialects
Quoted identifier preservation and cross-dialect conversion ("id" ↔ `id` ↔ [id])
Optimize SQL queries (constant folding, boolean simplification, subquery unnesting, predicate pushdown, column qualification)
Scope analysis — track query scopes, sources, columns, and correlations across nested queries
Schema management — dialect-aware table/column registration and type lookup
AST traversal — walk, find, transform expressions
AST Diff — semantic comparison of SQL statements with structural change detection
SQL Execution Engine — in-memory query execution against Rust data structures for testing and validation
CTEs, subqueries, set operations (UNION / INTERSECT / EXCEPT)
Window functions with frames and filters
CAST, TRY_CAST, EXTRACT, INTERVAL, EXISTS, ANY/ALL/SOME
Full DDL support (CREATE TABLE, ALTER TABLE, DROP TABLE, CREATE VIEW, etc.)
Serde serialization for AST nodes

Supported Dialects

All 30 dialects from Python's sqlglot are supported, with function name mapping, data type mapping, and ILIKE rewriting between dialects.

Official Dialects

Dialect	Description
ANSI SQL	Base SQL standard
Athena	AWS Athena (Presto-based)
BigQuery	Google BigQuery
ClickHouse	ClickHouse
Databricks	Databricks (Spark-based)
DuckDB	DuckDB
Hive	Apache Hive
MySQL	MySQL
Oracle	Oracle Database
PostgreSQL	PostgreSQL
Presto	Presto
Redshift	Amazon Redshift (Postgres-based)
Snowflake	Snowflake
Spark	Apache Spark SQL
SQLite	SQLite
StarRocks	StarRocks (MySQL-compatible)
Trino	Trino (Presto successor)
T-SQL	Microsoft SQL Server

Community Dialects

Dialect	Description
Doris	Apache Doris (MySQL-compatible)
Dremio	Dremio
Drill	Apache Drill
Druid	Apache Druid
Exasol	Exasol
Fabric	Microsoft Fabric (T-SQL variant)
Materialize	Materialize (Postgres-compatible)
PRQL	Pipelined Relational Query Language
RisingWave	RisingWave (Postgres-compatible)
SingleStore	SingleStore (MySQL-compatible)
Tableau	Tableau
Teradata	Teradata

Dialect Transform Rules

The transpiler applies dialect-specific rewrite rules when converting between dialects:

Rule	Example
Function name mapping	`NOW()` → `CURRENT_TIMESTAMP()`, `GETDATE()`
`SUBSTR` ↔ `SUBSTRING`	Postgres uses `SUBSTRING`, MySQL uses `SUBSTR`
`CEIL` → `CEILING`	T-SQL uses `CEILING`
`POW` → `POWER`	T-SQL/Oracle use `POWER`
`DATE_TRUNC` → `DATETRUNC`/`TRUNC`	T-SQL uses `DATETRUNC`, Oracle uses `TRUNC`
`ARRAY_AGG` → `LIST`/`COLLECT_LIST`	DuckDB uses `LIST`, Hive/Spark use `COLLECT_LIST`
`HEX`/`UNHEX` → `TO_HEX`/`FROM_HEX`	Presto/Trino naming convention
`IFNULL` → `COALESCE`	MySQL `IFNULL` → ANSI `COALESCE`
`IFNULL` → `ISNULL`	MySQL `IFNULL` → T-SQL `ISNULL`
`NVL` → `COALESCE`	Oracle `NVL` → standard `COALESCE`
`LEN` ↔ `LENGTH`	T-SQL `LEN` ↔ standard `LENGTH`
`RAND` ↔ `RANDOM`	MySQL `RAND` ↔ Postgres `RANDOM`
`ILIKE` → `LOWER`/`LIKE`	Rewritten for dialects without native ILIKE
`LIMIT` ↔ `TOP`	`LIMIT 10` → `TOP 10` for T-SQL
`LIMIT` → `FETCH FIRST`	`LIMIT 10` → `FETCH FIRST 10 ROWS ONLY` (Oracle)
Quoted identifiers	`"id"` ↔ `id` ↔ `[id]` per dialect
Data type mapping	`TEXT` ↔ `STRING`, `INT` → `BIGINT` (BigQuery)
`BYTEA` ↔ `BLOB`	Postgres `BYTEA` ↔ MySQL `BLOB`

Recent Parser Improvements and Benchmark Snapshot

This repository has recently added tolerance and compatibility fixes across multiple dialects, including:

ClickHouse: richer dotted/typed access parsing (expr.:Type, expr.^field, .null field paths), plus broader support for alias-heavy query forms.
DuckDB: alias-first projection support (alias: expr) and DESCRIBE-in-subquery tolerance.
PostgreSQL: BIT VARYING(n) cast tolerance, JSON key/value argument parsing ('k' : v), richer ON CONFLICT (...) target parsing, and broader geometric/operator-sequence tolerance.
MySQL: improved UPDATE/DELETE tolerance around ORDER BY, LIMIT, PARTITION, and insert alias edge cases.
T-SQL / ANSI extensions: tolerance for WITH XMLNAMESPACES (...) and additional parser guardrails for mixed corpora.

Benchmark reference:

SQL AST Benchmark

Acceptance Report (latest run)

dialect	total	accepted	rejected	accept%
postgresql	29402	28089	1313	95.53%
sqlite	12119	12103	16	99.87%
mysql	30220	29925	295	99.02%
clickhouse	92268	91148	1120	98.79%
duckdb	41148	40029	1119	97.28%
hive	41294	40774	520	98.74%
spark_sql	14464	14180	284	98.04%
trino	71	70	1	98.59%
tsql	14782	13521	1261	91.47%
oracle	21648	21608	40	99.82%
bigquery	224	222	2	99.11%
redshift	2992	2964	28	99.06%
multi	10962	9637	1325	87.91%
TOTAL	311594	304270	7324	97.65%

Brief Summary of Not-Yet-Supported Query Kinds

The remaining rejects are concentrated in a few recurring categories:

Non-SQL or shell-like corpus lines mixed into SQL datasets (for example raw text, script fragments, malformed separators).
Dialect-specific meta commands or client commands (for example backslash-style command lines in PostgreSQL corpora).
Template or parameterized placeholders that are not concrete SQL syntax until preprocessed.
Highly dialect-specific operator families and parser corner-cases in advanced analytical or geometric expressions.
Intentionally malformed or truncated statements in test corpora (unterminated strings/comments, unexpected EOF).

Quick Start

Add to your Cargo.toml:

[dependencies]
sqlglot-rust = "0.10.29"

Parse and generate SQL

use sqlglot_rust::{parse, generate,Dialect};fnmain(){// Parse a SQL querylet ast = parse("SELECT a, b FROM t WHERE a > 1",Dialect::Ansi).unwrap();// Generate SQL (roundtrip)let sql = generate(&ast,Dialect::Ansi);assert_eq!(sql,"SELECT a, b FROM t WHERE a > 1");}

Transpile between dialects

use sqlglot_rust::{transpile,Dialect};fnmain(){// Postgres → BigQuery: NOW() becomes CURRENT_TIMESTAMP()let sql = transpile("SELECT NOW(), SUBSTRING(name, 1, 3) FROM users",Dialect::Postgres,Dialect::BigQuery,).unwrap();// → "SELECT CURRENT_TIMESTAMP(), SUBSTRING(name, 1, 3) FROM users"// Oracle → T-SQL: NVL becomes ISNULLlet sql = transpile("SELECT NVL(a, b) FROM t",Dialect::Oracle,Dialect::Tsql,).unwrap();// → "SELECT ISNULL(a, b) FROM t"}

Parse a dialect name from a string

use sqlglot_rust::Dialect;let d = Dialect::from_str("postgres").unwrap();assert_eq!(d,Dialect::Postgres);// Multiple aliases are supportedassert_eq!(Dialect::from_str("tsql"),Some(Dialect::Tsql));assert_eq!(Dialect::from_str("mssql"),Some(Dialect::Tsql));assert_eq!(Dialect::from_str("sqlserver"),Some(Dialect::Tsql));

Build SQL with the Expression Builder API

use sqlglot_rust::builder::{select, condition, column, literal};use sqlglot_rust::{generate,Dialect};fnmain(){// Fluent SELECT builderlet query = select(&["u.id","u.name","COUNT(o.id) AS order_count"]).from("users").left_join("orders","u.id = o.user_id").where_clause("u.active = true").and_where("o.created_at > '2024-01-01'").group_by(&["u.id","u.name"]).having("COUNT(o.id) > 0").order_by(&["order_count DESC"]).limit(10).build();let sql = generate(&query,Dialect::Postgres);// => "SELECT u.id, u.name, COUNT(o.id) AS order_count FROM users LEFT JOIN orders ON u.id = o.user_id WHERE u.active = true AND o.created_at > '2024-01-01' GROUP BY u.id, u.name HAVING COUNT(o.id) > 0 ORDER BY order_count DESC LIMIT 10"// Condition builderlet cond = condition("x = 1").and("y = 2").or("z = 3").build();// Expression factory functionslet col = column("name",Some("users"));// users.namelet num = literal(42);// 42}

Supported Statements

SELECT (with JOINs, WHERE, GROUP BY, HAVING, ORDER BY, LIMIT, OFFSET, TOP, FETCH FIRST, QUALIFY)
INSERT INTO ... VALUES / INSERT INTO ... SELECT
UPDATE ... SET ... WHERE (with RETURNING)
DELETE FROM ... WHERE (with USING)
MERGE INTO ... USING ... WHEN MATCHED / WHEN NOT MATCHED (with dialect extensions)
CREATE TABLE (with constraints, IF NOT EXISTS, TEMPORARY, AS SELECT)
DROP TABLE (with IF EXISTS)
ALTER TABLE (ADD COLUMN, DROP COLUMN, RENAME COLUMN, RENAME TO, ALTER TYPE)
CREATE VIEW / DROP VIEW
TRUNCATE TABLE
BEGIN / COMMIT / ROLLBACK
EXPLAIN
USE

Expression Support

Column references (qualified and unqualified) with quote-style metadata
Numeric, string, boolean, and NULL literals
Binary operators (+, -, *, /, %, =, <>, <, >, <=, >=, AND, OR, ||)
Unary operators (NOT, -, +)
Bitwise operators (&, |, ^, <<, >>)
Function calls (with DISTINCT, FILTER, OVER support)
Typed function expressions (72+ functions: date/time, string, aggregate, array, JSON, window, math, conversion)
BETWEEN, IN, IS NULL, LIKE, ILIKE
CASE ... WHEN ... THEN ... ELSE ... END
CAST, TRY_CAST, PostgreSQL :: cast
EXTRACT(field FROM expr) for all date/time fields
INTERVAL expressions
EXISTS, IN (subquery)
Array literals (ARRAY[1, 2, 3])
Window functions with frames (ROWS, RANGE, GROUPS)
Common Table Expressions (WITH / WITH RECURSIVE)
Set operations (UNION, INTERSECT, EXCEPT with ALL)
Parenthesized sub-expressions and subqueries

CLI

The CLI is built with the cli feature flag and provides three commands:

Install

cargo install sqlglot-rust --features cli

Transpile

Transpile SQL between dialects:

# From stdinecho"SELECT CAST(x AS INT) FROM t"| sqlglot transpile --read mysql --write postgres
# With pretty-printingecho"SELECT a, b FROM t WHERE x > 1"| sqlglot transpile --pretty
# With optimizerecho"SELECT * FROM t WHERE 1 = 1 AND a > 5"| sqlglot transpile --optimize
# From file to file
sqlglot transpile --read mysql --write postgres --input query.sql --output result.sql

Parse

Parse SQL and output the AST as JSON:

echo"SELECT a FROM t"| sqlglot parse --pretty

Format

Pretty-print SQL:

echo"select a,b from t where x>1"| sqlglot format

Options

Option	Description
`--read <dialect>`	Source dialect (default: ansi)
`--write <dialect>`	Target dialect (default: ansi)
`--pretty`	Pretty-print output
`--input <file>`	Read from file instead of stdin
`--output <file>`	Write to file instead of stdout
`--optimize`	Run optimizer before generation (transpile only)

C/C++ FFI

sqlglot-rust can be built as a shared library (.so / .dylib / .dll) or a static library (.a) for use from C, C++, or any language that supports the C ABI.

Build the FFI libraries

# Build for the current host
make ffi
# Build for a specific target
make ffi-macos-arm64 # aarch64-apple-darwin
make ffi-macos-amd64 # x86_64-apple-darwin
make ffi-linux-amd64 # x86_64-unknown-linux-gnu
make ffi-linux-arm64 # aarch64-unknown-linux-gnu# Build all four targets
make ffi-all

Output goes to target/ffi/ with the C header at target/ffi/include/sqlglot.h.

C API

#include"sqlglot.h"constchar*sqlglot_version(void);
char*sqlglot_parse(constchar*sql, constchar*dialect);
char*sqlglot_transpile(constchar*sql, constchar*from_dialect, constchar*to_dialect);
char*sqlglot_generate(constchar*ast_json, constchar*dialect);
char*sqlglot_qualify_columns(constchar*ast_json, constchar*schema_json, constchar*dialect);
char*sqlglot_build_scope(constchar*ast_json);
char*sqlglot_lineage(
constchar*column,
constchar*ast_json,
constchar*schema_json,
constchar*config_json
);
voidsqlglot_free(char*ptr); /* releases an owned API result; NULL-safe */

Schema-aware calls accept an ordered table-path transport:

{
"tables": [
{
"path": ["public", "cards"],
"columns": [
{"name": "id", "data_type": "BIGINT"},
{"name": "amount", "data_type": "DECIMAL(18, 4)"}
]
}
]
}

Identifier path components are explicit, column order is retained for wildcard expansion, and data_type uses the parser's SQL data-type grammar.

C usage example

#include<stdio.h>#include"sqlglot.h"intmain(void) {
char*result=sqlglot_transpile("SELECT * FROM t LIMIT 10", "mysql", "tsql");
if (result) {
printf("%s\n", result); /* → SELECT TOP 10 * FROM t */sqlglot_free(result);
}
return0;
}

See examples/ffi_example.c and examples/ffi_example.cpp for complete examples.

Link against the library

# macOS
gcc example.c -Itarget/ffi/include -Ltarget/release -lsqlglot_rust -o example
# Linux
gcc example.c -Itarget/ffi/include -Ltarget/release -lsqlglot_rust -lpthread -ldl -lm -o example
LD_LIBRARY_PATH=target/release ./example

Documentation

Installation — Dependency setup and verification
Developer Guide — Parsing, generating, transpiling, AST traversal, optimization, and serialization with full code examples
API Reference — Complete type catalog, function signatures, dialect tables, and error variants

Architecture

src/
├── ast/ # AST node definitions (~40 expression types, 15 statement types)
├── bin/ # CLI binary (sqlglot) — feature-gated behind "cli"
├── builder/ # Expression builder API (fluent SQL construction)
├── dialects/ # 30 dialect definitions with transform rules
├── diff/ # AST diff / semantic SQL comparison
├── errors/ # Error types
├── executor/ # In-memory SQL execution engine
├── ffi.rs # C-compatible FFI bindings (extern "C" API)
├── generator/ # SQL code generator
├── lib.rs # Public API (parse, generate, transpile)
├── optimizer/ # Query optimization and scope analysis
├── parser/ # Recursive-descent SQL parser
├── planner/ # Logical query planner (execution plan DAG)
├── schema/ # Schema management (MappingSchema, dialect-aware lookups)
└── tokens/ # Token types (~200+ variants) and tokenizer

Development

# Build
cargo build
# Build with CLI
cargo build --features cli
# Run tests (500+ tests)
cargo test# Run CLI tests
cargo test --features cli --test test_cli
# Run benchmarks
cargo bench
# Lint
cargo clippy
# Format
cargo fmt

A Makefile is provided for convenience:

make build # cargo build
make test# cargo test
make lint # cargo clippy
make fmt # cargo fmt
make sbom # Generate SPDX SBOM (see below)
make bump-version # Update version across all files (see below)
make ffi # Build FFI libraries for the current host
make ffi-all # Build FFI libraries for all 4 targets (macOS/Linux × arm64/amd64)
make all # build + sbom

SBOM (Software Bill of Materials)

The project supports generating an SBOM in SPDX 2.3 JSON format using cargo-sbom.

Prerequisites

Install cargo-sbom:

cargo install cargo-sbom

Generate the SBOM

make sbom

This writes the SBOM to target/sbom/sqlglot-rust.spdx.json. You can also run the command directly:

cargo sbom --output-format spdx_json_2_3 > target/sbom/sqlglot-rust.spdx.json

The generated SBOM includes all dependency packages with license information, download locations, and Package URLs (PURLs).

Updating the Version

Use the bump-version Makefile target to update the version consistently across all configuration and documentation files:

make bump-version VERSION=1.0.0

This updates:

Cargo.toml — package version
README.md — dependency snippet
docs/installation.md — dependency snippet
Cargo.lock — regenerated automatically

The VERSION parameter is required and must be a full semantic version (e.g. 1.0.0, 0.10.1).

Acknowledgements

This project is inspired by and derived from sqlglot by Toby Mao, licensed under the MIT License. The original sqlglot license is included in this repository as required by the MIT License terms. A copy of the original license can also be found in the polyglot repository.

License

This project is licensed under the MIT License — see the LICENSE file for details.

This project includes code and concepts derived from sqlglot, which is also licensed under the MIT License. The original copyright notice and permission notice are reproduced below:

MIT License
Copyright (c) 2023 Toby Mao
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
benches		benches
docs		docs
examples		examples
packaging		packaging
scripts		scripts
src		src
tests		tests
.dockerignore		.dockerignore
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
Cargo.lock		Cargo.lock
Cargo.toml		Cargo.toml
LICENSE		LICENSE
Makefile		Makefile
README.md		README.md
cbindgen.toml		cbindgen.toml

Uh oh!

Latest commit

History

Folders and files

Repository files navigation