feat(materialize): add new backend for Materialize streaming database

This commit adds comprehensive support for Materialize, including:
- Core backend implementation with cluster and connection management
- Support for materialized views, indexes, sources, sinks, and secrets
- Idiomatic Materialize SQL patterns (mz_now(), SUBSCRIBE, etc.)
- Test coverage for Materialize-specific features
- Integration with existing Ibis test suite

Author: jasonhernandez

.github/workflows/ibis-backends.yml

@@ -166,6 +166,13 @@ jobs:
             sys-deps:
               - libgeos-dev
               - libpq5
+          - name: materialize
+            title: Materialize
+            serial: true
+            services:
+              - materialize
+            extras:
+              - --extra materialize
           - name: risingwave
             title: RisingWave
             serial: true
@@ -335,6 +342,15 @@ jobs:
               sys-deps:
                 - libgeos-dev
                 - libpq5
+          - os: windows-latest
+            backend:
+              name: materialize
+              title: Materialize
+              serial: true
+              services:
+                - materialize
+              extras:
+                - --extra materialize
           - os: windows-latest
             backend:
               name: risingwave

README.md

@@ -135,7 +135,7 @@ This allows you to combine the flexibility of Python with the scale and performa
 
 ## Backends
 
-Ibis supports nearly 20 backends:
+Ibis supports 20 backends:
 
 - [Apache DataFusion](https://ibis-project.org/backends/datafusion/)
 - [Apache Druid](https://ibis-project.org/backends/druid/)
@@ -146,6 +146,7 @@ Ibis supports nearly 20 backends:
 - [ClickHouse](https://ibis-project.org/backends/clickhouse/)
 - [DuckDB](https://ibis-project.org/backends/duckdb/)
 - [Exasol](https://ibis-project.org/backends/exasol)
+- [Materialize](https://ibis-project.org/backends/materialize/)
 - [MySQL](https://ibis-project.org/backends/mysql/)
 - [Oracle](https://ibis-project.org/backends/oracle/)
 - [Polars](https://ibis-project.org/backends/polars/)

ci/schema/materialize.sql

@@ -0,0 +1,185 @@
+DROP TABLE IF EXISTS diamonds CASCADE;
+
+CREATE TABLE diamonds (
+    carat FLOAT,
+    cut TEXT,
+    color TEXT,
+    clarity TEXT,
+    depth FLOAT,
+    "table" FLOAT,
+    price BIGINT,
+    x FLOAT,
+    y FLOAT,
+    z FLOAT
+);
+
+-- Note: In real usage, data would be loaded via sources or INSERT statements
+-- For CI, we'll need to mount CSV files and use a different loading strategy
+
+DROP TABLE IF EXISTS astronauts CASCADE;
+
+CREATE TABLE astronauts (
+    "id" BIGINT,
+    "number" BIGINT,
+    "nationwide_number" BIGINT,
+    "name" VARCHAR,
+    "original_name" VARCHAR,
+    "sex" VARCHAR,
+    "year_of_birth" BIGINT,
+    "nationality" VARCHAR,
+    "military_civilian" VARCHAR,
+    "selection" VARCHAR,
+    "year_of_selection" BIGINT,
+    "mission_number" BIGINT,
+    "total_number_of_missions" BIGINT,
+    "occupation" VARCHAR,
+    "year_of_mission" BIGINT,
+    "mission_title" VARCHAR,
+    "ascend_shuttle" VARCHAR,
+    "in_orbit" VARCHAR,
+    "descend_shuttle" VARCHAR,
+    "hours_mission" DOUBLE PRECISION,
+    "total_hrs_sum" DOUBLE PRECISION,
+    "field21" BIGINT,
+    "eva_hrs_mission" DOUBLE PRECISION,
+    "total_eva_hrs" DOUBLE PRECISION
+);
+
+DROP TABLE IF EXISTS batting CASCADE;
+
+CREATE TABLE batting (
+    "playerID" TEXT,
+    "yearID" BIGINT,
+    "stint" BIGINT,
+    "teamID" TEXT,
+    "lgID" TEXT,
+    "G" BIGINT,
+    "AB" BIGINT,
+    "R" BIGINT,
+    "H" BIGINT,
+    "X2B" BIGINT,
+    "X3B" BIGINT,
+    "HR" BIGINT,
+    "RBI" BIGINT,
+    "SB" BIGINT,
+    "CS" BIGINT,
+    "BB" BIGINT,
+    "SO" BIGINT,
+    "IBB" BIGINT,
+    "HBP" BIGINT,
+    "SH" BIGINT,
+    "SF" BIGINT,
+    "GIDP" BIGINT
+);
+
+DROP TABLE IF EXISTS awards_players CASCADE;
+
+CREATE TABLE awards_players (
+    "playerID" TEXT,
+    "awardID" TEXT,
+    "yearID" BIGINT,
+    "lgID" TEXT,
+    "tie" TEXT,
+    "notes" TEXT
+);
+
+DROP TABLE IF EXISTS functional_alltypes CASCADE;
+
+CREATE TABLE functional_alltypes (
+    "id" INTEGER,
+    "bool_col" BOOLEAN,
+    "tinyint_col" SMALLINT,
+    "smallint_col" SMALLINT,
+    "int_col" INTEGER,
+    "bigint_col" BIGINT,
+    "float_col" REAL,
+    "double_col" DOUBLE PRECISION,
+    "date_string_col" TEXT,
+    "string_col" TEXT,
+    "timestamp_col" TIMESTAMP WITHOUT TIME ZONE,
+    "year" INTEGER,
+    "month" INTEGER
+);
+
+DROP TABLE IF EXISTS tzone CASCADE;
+
+CREATE TABLE tzone (
+    "ts" TIMESTAMP WITH TIME ZONE,
+    "key" TEXT,
+    "value" DOUBLE PRECISION
+);
+
+INSERT INTO tzone
+    SELECT
+	CAST('2017-05-28 11:01:31.000400' AS TIMESTAMP WITH TIME ZONE) +
+	    (t * INTERVAL '1 day' + t * INTERVAL '1 second') AS "ts",
+	CHR(97 + t) AS "key",
+	t + t / 10.0 AS "value"
+    FROM generate_series(0, 9) AS t;
+
+DROP TABLE IF EXISTS array_types CASCADE;
+
+CREATE TABLE IF NOT EXISTS array_types (
+    "x" BIGINT[],
+    "y" TEXT[],
+    "z" DOUBLE PRECISION[],
+    "grouper" TEXT,
+    "scalar_column" DOUBLE PRECISION,
+    "multi_dim" BIGINT[][]
+);
+
+-- Note: Materialize does not currently support multi-dimensional arrays containing NULL sub-arrays, so they are not included.
+-- Fix pending: https://github.com/MaterializeInc/materialize/pull/33786
+
+INSERT INTO array_types VALUES
+    (ARRAY[1::BIGINT, 2::BIGINT, 3::BIGINT], ARRAY['a', 'b', 'c'], ARRAY[1.0::DOUBLE PRECISION, 2.0::DOUBLE PRECISION, 3.0::DOUBLE PRECISION], 'a', 1.0, NULL),
+    (ARRAY[4::BIGINT, 5::BIGINT], ARRAY['d', 'e'], ARRAY[4.0::DOUBLE PRECISION, 5.0::DOUBLE PRECISION], 'a', 2.0, NULL),
+    (ARRAY[6::BIGINT, NULL], ARRAY['f', NULL], ARRAY[6.0::DOUBLE PRECISION, NULL::DOUBLE PRECISION], 'a', 3.0, NULL),
+    (ARRAY[NULL::BIGINT, 1::BIGINT, NULL::BIGINT], ARRAY[NULL, 'a', NULL], ARRAY[]::DOUBLE PRECISION[], 'b', 4.0, NULL),
+    (ARRAY[2::BIGINT, NULL, 3::BIGINT], ARRAY['b', NULL, 'c'], NULL, 'b', 5.0, NULL),
+    (ARRAY[4::BIGINT, NULL, NULL, 5::BIGINT], ARRAY['d', NULL, NULL, 'e'], ARRAY[4.0::DOUBLE PRECISION, NULL::DOUBLE PRECISION, NULL::DOUBLE PRECISION, 5.0::DOUBLE PRECISION], 'c', 6.0, NULL);
+
+DROP TABLE IF EXISTS json_t CASCADE;
+
+CREATE TABLE IF NOT EXISTS json_t (rowid BIGINT, "js" JSONB);
+
+INSERT INTO json_t VALUES
+    (1, '{"a": [1,2,3,4], "b": 1}'),
+    (2, '{"a":null,"b":2}'),
+    (3, '{"a":"foo", "c":null}'),
+    (4, 'null'),
+    (5, '[42,47,55]'),
+    (6, '[]'),
+    (7, '"a"'),
+    (8, '""'),
+    (9, '"b"'),
+    (10, NULL),
+    (11, 'true'),
+    (12, 'false'),
+    (13, '42'),
+    (14, '37.37');
+
+DROP TABLE IF EXISTS win CASCADE;
+
+CREATE TABLE win ("g" TEXT, "x" BIGINT, "y" BIGINT);
+
+INSERT INTO win VALUES
+    ('a', 0, 3),
+    ('a', 1, 2),
+    ('a', 2, 0),
+    ('a', 3, 1),
+    ('a', 4, 1);
+
+DROP TABLE IF EXISTS topk CASCADE;
+
+CREATE TABLE topk ("x" BIGINT);
+
+INSERT INTO topk VALUES (1), (1), (NULL);
+
+DROP TABLE IF EXISTS map CASCADE;
+
+CREATE TABLE map (idx BIGINT, kv JSONB);
+
+INSERT INTO map VALUES
+    (1, '{"a": 1, "b": 2, "c": 3}'),
+    (2, '{"d": 4, "e": 5, "f": 6}');

compose.yaml

@@ -556,6 +556,28 @@ services:
     networks:
       - impala
 
+  materialize:
+    image: materialize/materialized:v0.160.0
+    ports:
+      - 6875:6875 # SQL port
+      - 6876:6876 # HTTP port
+      - 6874:6874 # Internal port
+    environment:
+      MZ_LOG: info
+      MZ_TELEMETRY: "false"
+      MZ_SYSTEM_PARAMETER_DEFAULT: "max_tables=1000" # Increase from default 200 for testing
+    healthcheck:
+      test:
+        - CMD-SHELL
+        - bash -c 'printf "GET / HTTP/1.1\n\n" > /dev/tcp/127.0.0.1/6875; exit $$?;'
+      interval: 2s
+      retries: 30
+      timeout: 5s
+    volumes:
+      - materialize:/data
+    networks:
+      - materialize
+
   risingwave:
     image: ghcr.io/risingwavelabs/risingwave:v2.1.3
     command: "standalone --meta-opts=\" \
@@ -623,6 +645,7 @@ networks:
   oracle:
   exasol:
   flink:
+  materialize:
   risingwave:
   spark-connect:
 
@@ -635,5 +658,6 @@ volumes:
   postgres:
   exasol:
   impala:
+  materialize:
   risingwave:
   spark-connect:

docs/backends/materialize.qmd

@@ -0,0 +1,69 @@
+# Materialize
+
+[https://materialize.com](https://materialize.com)
+
+![](https://img.shields.io/badge/memtables-fallback-yellow?style=flat-square) ![](https://img.shields.io/badge/inputs-Materialize tables | Streaming sources-blue?style=flat-square) ![](https://img.shields.io/badge/outputs-Materialize tables | CSV | pandas | Parquet | PyArrow-orange?style=flat-square) ![](https://img.shields.io/badge/streaming-SUBSCRIBE-purple?style=flat-square)
+
+## Install
+
+Install Ibis and dependencies for the Materialize backend:
+
+```{.bash}
+pip install 'ibis-framework[materialize]'
+```
+
+And connect:
+
+```{.python}
+import ibis
+
+con = ibis.materialize.connect()  # <1>
+```
+
+1. Adjust connection parameters as needed.
+
+## Connect
+
+### `ibis.materialize.connect`
+
+```python
+con = ibis.materialize.connect(
+    user="materialize",
+    password="password",
+    host="localhost",
+    port=6875,
+    database="materialize",
+    cluster="quickstart",  # Optional: specify default cluster
+)
+```
+
+::: {.callout-note}
+`ibis.materialize.connect` is a thin wrapper around
+[`ibis.backends.materialize.Backend.do_connect`](#ibis.backends.materialize.Backend.do_connect).
+:::
+
+### Connection Parameters
+
+```{python}
+#| echo: false
+#| output: asis
+from _utils import render_do_connect
+
+render_do_connect("materialize")
+```
+
+### `ibis.connect` URL format
+
+In addition to `ibis.materialize.connect`, you can also connect to Materialize by
+passing a properly-formatted connection URL to `ibis.connect`:
+
+```python
+con = ibis.connect(f"materialize://{user}:{password}@{host}:{port}/{database}")
+```
+
+```{python}
+#| echo: false
+BACKEND = "Materialize"
+```
+
+{{< include ./_templates/api.qmd >}}