ClickHouse SQL Parser for Java
General SQL Parser (GSP) parses ClickHouse with a dedicated, hand-tuned grammar — full AST access, offline syntax validation, formatting, rewriting, and column-level data lineage. One commercial SDK — ClickHouse is available in the Java edition.
ClickHouse SQL that GSP parses
CREATE TABLE hits (dt DateTime, uid UInt64)
ENGINE = MergeTree()
PARTITION BY toYYYYMM(dt)
ORDER BY (dt, uid); Need us to check your own ClickHouse SQL? Send a sample.
Parse it in Java
import gudusoft.gsqlparser.*;
TGSqlParser parser = new TGSqlParser(EDbVendor.dbvclickhouse);
parser.sqltext = sql; // the ClickHouse SQL above
if (parser.parse() == 0) {
System.out.println(parser.sqlstatements.size() + " statement(s) parsed");
} else {
System.out.println(parser.getErrormessage());
} Java edition only. The .NET library has no ClickHouse grammar, so this dialect cannot be parsed from C# or VB.NET. See the .NET SQL Parser SDK for the dialects it does cover.
What GSP handles in ClickHouse
- ENGINE = MergeTree clauses
- PREWHERE filtering
- ARRAY JOIN
- LIMIT BY and SAMPLE
- Materialized views ON CLUSTER
Beyond parsing, the same AST powers table/column extraction, validation, formatting, and column-level lineage for ClickHouse.
Recent ClickHouse parser updates
- 2026-08-18 [ClickHouse] Parse Map literals, two-argument CAST, dotted nested column names, and tuple element access (a.2)
- 2026-08-18 [ClickHouse] Parse ALTER TABLE UPDATE/DELETE IN PARTITION, INSERT ... FORMAT with inline data, and PARALLEL WITH
- 2026-08-17 [ClickHouse] Parse == comparison, INTERSECT DISTINCT, GROUPING SETS, ASOF/ANY join order, WITH FILL STALENESS/INTERPOLATE, VALUES table literals
Full history in the release notes.
ClickHouse resources
- ClickHouse syntax reference — dialect documentation
- ClickHouse keyword compatibility — every keyword, and whether it can be an identifier
- ClickHouse capability data (JSON) — measured construct-by-construct parse coverage, machine-readable
- Java quick start — Maven/Gradle setup and first parse
Common questions
Do I need a live ClickHouse database connection to validate SQL?
No. GSP validates ClickHouse syntax completely offline. You get error line and column positions, the offending token, and a hint — with no server, driver, or credentials involved.
How complete is the ClickHouse grammar?
GSP uses a dedicated hand-tuned grammar for ClickHouse — not a generic SQL grammar with flags. The parser recognizes 554 ClickHouse keywords and knows which 551 of them can also be used as identifiers, which is exactly the kind of edge case that breaks generic parsers.
Can I use the ClickHouse parser from C# / .NET?
Not currently. ClickHouse is a Java-edition dialect — the .NET library ships 15 dedicated grammars and ClickHouse is not one of them, so there is no C# code path for it. Everything else on this page describes the Java SDK. If you need ClickHouse in .NET, email info@sqlparser.com; customer demand is how dialects get prioritized for the .NET port.
Can GSP extract column-level lineage from ClickHouse SQL?
Yes. The built-in DataFlowAnalyzer produces column-level lineage, impact analysis, and call graphs from ClickHouse scripts — it is the engine behind Gudu SQLFlow and the lineage integrations for DataHub and OpenMetadata.