Engine Cheatsheets
One page per database that answers, fast: how do I connect from Irodori, what is the query model, and what are the per-engine quirks I will trip on. These are the human-facing, copy-pasteable companion to the deeper engine syntax reference (driver/decoding internals) and data-source support status (coverage).
Each cheatsheet is meant to be generated from the local knowledge inputs in
irodori-table (knowledge/cheatsheets/*.json plus the knowledge DB). Until a
page is generated, it may be hand-seeded and marked <!-- seed -->. Generator
ownership and drift rules are tracked in
repository-boundaries.md.
Index
| Cheatsheet | Engine(s) covered | Status |
|---|---|---|
| neo4j.md | Neo4j (graph, Bolt/Cypher); Memgraph extension notes | Seed (flagship graph/Bolt page) |
| postgres.md | PostgreSQL (+ Cockroach/Yugabyte/Redshift/Timescale/Neon; H2 wire notes) | Generated (knowledge/cheatsheets/postgres.json) |
| questdb.md | QuestDB (Postgres wire + time-series SQL extensions) | Seed |
| mysql.md | MySQL / MariaDB / TiDB | Planned |
| sqlite.md | SQLite | Planned |
| oracle.md | Oracle | Planned |
| sqlserver.md | SQL Server | Planned |
| duckdb.md | DuckDB / MotherDuck | Planned |
| mongodb.md | MongoDB | Planned |
| redis.md | Redis | Planned |
| cassandra.md | Cassandra / ScyllaDB | Planned |
| clickhouse.md | ClickHouse | Planned |
| snowflake.md | Snowflake | Planned |
| bigquery.md | BigQuery | Planned |
| bigtable.md | Bigtable | Planned |
| influxdb.md | InfluxDB | Planned |
New cheatsheets are added only for engines that are at least Wired in
data-source-support-status.md. An engine
that is “Recognized, extension required”
or “Not registered” gets a row in the support-status doc, not a cheatsheet, until
it can actually connect.
Maintenance queue
Seed the next pages in the same order as the support-status table, prioritizing
engines with verified or wired query paths and enough source coverage in
knowledge/sources.json: duckdb.md, mongodb.md, redis.md,
cassandra.md, clickhouse.md, snowflake.md, bigquery.md, bigtable.md,
and influxdb.md.
Some sibling connector implementations can run ahead of the root registry. Until the registry and support-status page promote an engine to Wired or Extension, keep those details as related notes under the nearest wired cheatsheet instead of publishing a standalone page.
Page format (the template every cheatsheet follows)
Keep sections in this order so the generator can produce them deterministically and so readers build muscle memory:
- At a glance — wire/driver, default port, query language, Irodori support status, and a one-line “what makes this engine different”.
- Connect — Irodori connection fields and the raw URL/DSN form, with a minimal working example.
- Query model — what you type, what comes back, row cap behavior.
- Essential statements — a tight, runnable set of the 80%-case queries.
- Introspection — how to list objects the way Irodori’s object browser does.
- Irodori-specific behavior — mapping/quirks unique to how this app handles the engine (decoding, object-browser mapping, what’s not implemented yet).
- Gotchas — the small number of things that actually bite people.
- Sources — the
knowledge/sources.jsonids the page was generated from.
The Sources footer is load-bearing: it ties each page back to the official docs in the knowledge registry so a refresh can detect when a page is stale.