Cross-Database Data Type Mapping Reference

About this tool

Every database engine has its own type system, and the same logical concept — "a whole number", "some text", "a unique identifier" — often maps to a differently-named (or differently-behaved) column type in each one. This reference lists the closest real equivalent in five widely used dialects for common abstract types, including the quirks worth knowing: SQLite's type affinity system means it doesn't enforce fixed-width integer or text sizes the way the others do, several engines have no native UUID type and fall back to a fixed-length string or binary column, and "auto-increment primary key" is implemented completely differently across engines (dedicated keyword, sequence-backed serial type, identity column, or rowid alias).

The filter matches the general type, related words and the database-specific names, so typing bytea, datetimeoffset or nvarchar finds the row they belong to. Rows cover whole numbers of every size, decimals, text (fixed, short, long and Unicode), dates and times with and without a time zone, booleans, binary data, UUIDs, JSON, enumerated values, arrays and auto-increment keys.