sproutdb

SproutDB Reference Skill

SproutDB is a custom embedded/networked database with its own query language (NOT SQL). Never guess syntax — always check the reference.

When to read which file

Task	Read
Writing queries (get, upsert, delete, create table, etc.)	references/query-language.md
Designing tables, choosing types, modeling relationships	references/schema-design.md
C# integration (DI, migrations, ISproutDatabase, responses)	references/csharp-integration.md

Critical differences from SQL

• UPSERT, not INSERT/UPDATE — upsert table {key: 'val'} with JSON-like body
• GET, not SELECT — get table where ... not select * from table where ...
• FOLLOW, not JOIN — follow source.col -> target.col as alias
• No parentheses in column definitions — create table t (name string 100) not (name string(100))
• No semicolons — queries are single-line or multi-line without terminators
• Comments use ## — not -- or /* */
• String literals use single quotes — 'hello' not "hello"
• _id is automatic — ulong, auto-increment, cannot be set manually on insert
• No ALTER TABLE — use add column, rename column, alter column as separate commands
• DELETE requires WHERE — no way to delete all rows in one statement
• No foreign keys — relationships via follow at query time only
• No arrays/JSON — use junction tables instead