ClickHouse SQL Syntax for VS Code

A VS Code extension providing rich syntax highlighting, intelligent completions, hover documentation, and code formatting for ClickHouse SQL dialect.

Features

🎨 Syntax Highlighting

Full syntax highlighting for ClickHouse-specific constructs:

• ClickHouse data types: UInt8, UInt64, Float32, Float64, Decimal, String, FixedString, Date, DateTime, DateTime64, Array, Tuple, Map, Nested, Nullable, LowCardinality, IPv4, IPv6, UUID, and more
• Table engines: MergeTree, ReplicatedMergeTree, AggregatingMergeTree, CollapsingMergeTree, VersionedCollapsingMergeTree, Distributed, Kafka, S3, and more
• ClickHouse-specific keywords: PREWHERE, SAMPLE, FINAL, GLOBAL IN, ARRAY JOIN, WITH TOTALS, WITH ROLLUP, WITH CUBE, FORMAT, SETTINGS, TTL, CODEC, and more
• ClickHouse functions: 400+ functions across categories (aggregate, array, string, date/time, math, type conversion, conditional)
• Standard SQL keywords, operators, string literals, numbers, and comments

✍️ Code Snippets

40+ ready-to-use snippets for common ClickHouse patterns:

📖 Hover Documentation

Hover over ClickHouse functions to see:

• Function signature
• Description
• Usage examples

💡 Auto-Completion

IntelliSense completions for:

• All ClickHouse keywords and clauses
• 400+ ClickHouse functions with snippet templates
• ClickHouse data types
• Available for both .chsql and .sql files

🔧 SQL Formatter

The formatter completely restructures your ClickHouse SQL for maximum readability:

• Each SQL clause on its own line — SELECT, FROM, WHERE, PREWHERE, GROUP BY, ORDER BY, HAVING, LIMIT, SETTINGS, FORMAT, all JOIN variants, UNION ALL, INTERSECT, EXCEPT
• Column lists expanded — every column in SELECT, GROUP BY, ORDER BY gets its own indented line
• Conditions expanded — AND / OR in WHERE, PREWHERE, HAVING, and JOIN ON each start a new indented line
• Subqueries stay nested — expressions inside parentheses are correctly skipped by the clause-splitter
• Literals protected — single-quoted strings, backtick/double-quoted identifiers, -- comments, and /* */ blocks are never altered
• Multi-statement support — multiple statements separated by ; are formatted independently with a blank line between them
• Three ways to invoke:
  • Right-click → ClickHouse: Format Document (context menu)
  • F1 → ClickHouse: Format Document (command palette)
  • ⇧⌥F — VS Code built-in Format Document shortcut

Before:

select user_id,count() as cnt,sum(revenue) as rev from events where event_date>=today()-30 and status='active' group by user_id order by cnt desc limit 100

After:

SELECT
    user_id,
    count() AS cnt,
    sum(revenue) AS rev
FROM events
WHERE
    event_date >= today() - 30
    AND status = 'active'
GROUP BY
    user_id
ORDER BY
    cnt DESC
LIMIT 100

Configurable options:

💉 SQL Injection

ClickHouse-specific syntax is also highlighted inside standard .sql files via grammar injection, so you get ClickHouse type and function highlighting without changing the file language.

File Association

The extension activates for files with extensions:

• .chsql — dedicated ClickHouse SQL files
• .ch.sql — ClickHouse SQL files with double extension

For standard .sql files, ClickHouse syntax is injected automatically via grammar injection.

You can also manually set the language to ClickHouse SQL using the language selector in the VS Code status bar.

Configuration

Examples

CREATE TABLE with MergeTree

CREATE TABLE analytics.events
(
    event_id     UInt64,
    user_id      UInt64,
    event_name   LowCardinality(String),
    event_time   DateTime,
    event_date   Date,
    properties   Map(String, String)
)
ENGINE = ReplicatedMergeTree('/clickhouse/tables/{shard}/analytics.events', '{replica}')
PARTITION BY toYYYYMM(event_date)
ORDER BY (user_id, event_time)
TTL event_date + INTERVAL 90 DAY
SETTINGS index_granularity = 8192;

Analytical Query with ClickHouse Functions

SELECT
    user_id,
    groupArray(10)(event_name)              AS last_events,
    uniqExact(session_id)                   AS sessions,
    quantile(0.95)(response_time_ms)        AS p95_response,
    windowFunnel(86400)(
        event_time,
        event_name = 'page_view',
        event_name = 'add_to_cart',
        event_name = 'purchase'
    )                                        AS funnel_step,
    dictGet('default.users', 'country', user_id) AS country
FROM analytics.events
WHERE event_date >= today() - 30
  AND event_name IN ('page_view', 'add_to_cart', 'purchase')
PREWHERE user_id > 0
GROUP BY user_id
HAVING funnel_step > 0
ORDER BY sessions DESC
LIMIT 1000
SETTINGS max_threads = 8, use_query_cache = 1
FORMAT JSONEachRow

Array Operations

SELECT
    arrayMap(x -> x * 2, [1, 2, 3, 4, 5])              AS doubled,
    arrayFilter(x -> x > 0, [-1, 0, 1, 2, 3])           AS positive,
    arraySort(arrayDistinct(groupArray(tag)))             AS unique_tags,
    arrayCumSum([1, 2, 3, 4, 5])                          AS running_total,
    arrayStringConcat(arrayMap(x -> toString(x), ids), ',') AS id_list
FROM events

Supported ClickHouse Versions

This extension supports ClickHouse 22.x and above, including all modern features:

• DateTime64 with sub-second precision
• Map data type
• JSON / Dynamic data types
• Variant type
• Window functions
• WITH TOTALS, WITH ROLLUP, WITH CUBE
• Named collections
• Refreshable Materialized Views

Contributing

Contributions are welcome! Please open issues or pull requests on GitHub.

License

MIT License