SQL

July 8, 2026 ยท View on GitHub

MongrelDB ships a DataFusion-backed SQL engine at POST /sql. From Zig, run SQL with Client.sql:

const rows = try db.sql(allocator, "SELECT 1");

This guide covers the SQL surface - DDL, DML, CREATE TABLE AS SELECT, recursive CTEs, and window functions - and when to reach for SQL versus the native query builder.

How sql behaves

db.sql(allocator, sql) sends {"sql": "..."} to /sql. It returns the decoded rows when the daemon replies with a JSON result set, and an empty slice with no error otherwise.

In practice:

  • DDL and DML (CREATE TABLE, INSERT, UPDATE, DELETE) reply with a non-JSON status body. sql returns an empty slice - success is the signal.
  • SELECT in most daemon builds streams Arrow IPC bytes rather than JSON. sql therefore returns an empty slice for SELECTs too. Use the native QueryBuilder for typed row retrieval in application code, and use sql for statements whose execution is the goal (DDL/DML/admin).

Errors are mapped to the same typed error set as everything else: an HTTP 400 or 5xx maps to error.Query/error.Http; 409 maps to error.Conflict; and so on. See errors.md.

db.sql(allocator, "INSERT INTO orders (id, customer, amount) VALUES (99, 'Zoe', 999.0)") catch |err| switch (err) {
    error.Conflict => std.debug.print("duplicate row\n", .{}),
    else => return err,
};

CREATE TABLE

Define a table in SQL instead of via Client.createTable. Column ids are assigned by the server when not stated.

_ = try db.sql(allocator,
    \\CREATE TABLE products (
    \\  id          INT64 PRIMARY KEY,
    \\  name        VARCHAR,
    \\  price       FLOAT64,
    \\  category    VARCHAR,
    \\  in_stock    BOOLEAN
    \\)
);

INSERT

_ = try db.sql(allocator, "INSERT INTO products (id, name, price, category, in_stock) VALUES (1, 'Widget', 9.99, 'tools', true)");
_ = try db.sql(allocator, "INSERT INTO products VALUES (2, 'Gadget', 19.99, 'tools', true)");

For bulk inserts, the native batch transaction (Client.begin) is usually faster because it stages ops in one round trip without re-parsing SQL.

UPDATE

_ = try db.sql(allocator, "UPDATE products SET price = 14.99 WHERE id = 1");
_ = try db.sql(allocator, "UPDATE orders SET amount = 200.0 WHERE customer = 'Bob'");

DELETE

_ = try db.sql(allocator, "DELETE FROM products WHERE in_stock = false");
_ = try db.sql(allocator, "DELETE FROM products WHERE id = 2");

SELECT

_ = try db.sql(allocator, "SELECT id, name FROM products WHERE category = 'tools' ORDER BY price");
_ = try db.sql(allocator, "SELECT category, COUNT(*) AS n FROM products GROUP BY category");

Remember SELECT bodies usually arrive as Arrow IPC, so sql returns an empty slice. To read rows back into Zig values, mirror the same lookup with the QueryBuilder.

CREATE TABLE AS SELECT

Materialize a query result into a new table. Great for snapshots, rollups, and denormalized aggregates.

// Snapshot all high-value orders into a new table.
_ = try db.sql(allocator, "CREATE TABLE archive AS SELECT * FROM orders WHERE amount > 500");

// Roll up sales by customer.
_ = try db.sql(allocator,
    \\CREATE TABLE sales_by_customer AS
    \\SELECT customer, SUM(amount) AS total
    \\FROM orders
    \\GROUP BY customer
);

The new table inherits column types from the query. Query it afterward with the native builder or SQL.

Recursive CTEs

WITH RECURSIVE is fully supported. Classic use cases: series generation, hierarchy/graph traversal.

// Generate the numbers 1..10.
_ = try db.sql(allocator,
    \\WITH RECURSIVE r(n) AS (
    \\  SELECT 1
    \\  UNION ALL
    \\  SELECT n + 1 FROM r WHERE n < 10
    \\)
    \\SELECT n FROM r
);

A common practical example is walking an adjacency list:

_ = try db.sql(allocator,
    \\WITH RECURSIVE descendants(id) AS (
    \\  SELECT id FROM categories WHERE id = 1
    \\  UNION ALL
    \\  SELECT c.id FROM categories c
    \\  JOIN descendants d ON c.parent_id = d.id
    \\)
    \\SELECT id FROM descendants
);

Window functions

Window functions compute aggregates/rankings across a moving window without collapsing rows. Useful for top-N-per-group, running totals, and row numbers.

// Row number within each customer, ordered by amount descending.
_ = try db.sql(allocator,
    \\SELECT id, customer, amount,
    \\       ROW_NUMBER() OVER (PARTITION BY customer ORDER BY amount DESC) AS rn
    \\FROM orders
);

// Running total per customer.
_ = try db.sql(allocator,
    \\SELECT id, customer, amount,
    \\       SUM(amount) OVER (PARTITION BY customer ORDER BY id) AS running_total
    \\FROM orders
);

RANK(), DENSE_RANK(), LAG(), LEAD(), NTILE(), and the usual window-frame clauses are available through DataFusion.

When to use SQL vs. the query builder

Both read from the same tables, but they are optimized for different jobs.

Reach forWhen
QueryBuilderPoint lookups, range scans, bitmap filters, full-text, and vector similarity that map to a native index. Sub-millisecond, no parser overhead, and rows decode into Zig values directly.
SQLDDL (CREATE TABLE, schemas, materialized views), multi-statement setup, joins, recursive CTEs, window functions, and arbitrary aggregates. Also the natural choice for admin scripts and one-off analysis.

Rules of thumb:

  • Need a typed slice of matching rows? Use the query builder.
  • Building/dropping tables, or running a CREATE TABLE AS SELECT? Use SQL.
  • Joining multiple tables, computing rankings, or walking a graph? Use SQL.
  • Filtering by one or more indexed columns? Use the query builder - it is faster and avoids Arrow-to-Zig decoding.

Mix freely: create tables with SQL, write rows with Client.put, read them back with QueryBuilder, and run analytics with SQL.

Next steps