SQL Commands

Singdata Lakehouse supports standard ANSI SQL, extended with incremental computation (Dynamic Table), change data capture (Table Stream), vector search, federated queries, and more. This manual covers the complete syntax reference for all SQL statements.

Migrating from another platform? See SQL Syntax Compatibility Reference for compatibility notes with Spark SQL / Databricks.

Query and Data Manipulation

Page	Description
Query Syntax	SELECT, JOIN, GROUP BY, window functions, Time Travel, EXPLAIN
Data Write and Modification	INSERT, UPDATE, DELETE, MERGE INTO, TRUNCATE
Data Import and Export	COPY INTO for file import, COPY INTO VOLUME for data export

Table and View Objects

Page	Description
Tables	Full DDL for regular tables: create, partition, bucket, primary key, ALTER, DROP
Views	Create, drop, and inspect views
Materialized Views	Create, refresh, alter, and drop materialized views
Dynamic Tables	Core incremental computation object: create, refresh configuration, ALTER, refresh history
Table Stream	CDC change capture object: create, inspect, drop
External Tables	Table definitions for external data sources such as Kafka external tables
Indexes	Create and manage bloom filter indexes, inverted indexes, and vector indexes

Storage, Connections, and Ecosystem

Page	Description
Schema	Create, alter, switch, and drop schemas
Volume	Create internal/external volumes, file operations (PUT/GET/LIST/REMOVE)
Pipe	Create and manage continuous data ingestion pipelines
Connection	Create and manage API, storage, and catalog connection objects
External Catalog & Schema	Federated queries: create and use external catalogs and external schemas
Synonym	Create and manage cross-schema object aliases
Data Sharing (Share)	Provider and consumer operations for cross-instance data sharing

Workspace and Access Management

Page	Description
Compute Cluster (VCluster)	Create, start/stop, configure, and drop compute clusters
Job Management	View, describe, and cancel running jobs
User Management	Add, modify, and remove workspace users
Roles and Privileges	Create roles and grant or revoke privileges
User-Defined Functions	Create and manage external functions and SQL functions

Basic Syntax

Page	Description
Basic Commands	Object identifier conventions, comment syntax, session parameters, general DDL patterns

Quick Reference

Object Naming Conventions

Singdata Lakehouse uses a three-part naming scheme: workspace_name.schema_name.object_name. The first part can be omitted when you are in the current workspace; the first two parts can be omitted when you are in the current schema.

General DDL Patterns

Most objects support four operations: CREATE [OR REPLACE] ... [IF NOT EXISTS], DROP ... [IF EXISTS], DESC[RIBE], and SHOW ...S [LIKE 'pattern'].

Document	Description
Object Model	Concepts, selection guidance, and how each object type works
INFORMATION_SCHEMA	Query metadata via SQL: table structure, job history, privilege assignments
SQL Functions	Complete built-in function reference