Supported data types with their categories and values - ApsaraDB for ClickHouse

ApsaraDB for ClickHouse supports the data types listed in this reference. For the complete upstream type system, see Data Types in the open-source ClickHouse documentation.

Supported data types

Integer

Signed integers. Use the smallest type that covers your value range to minimize storage.

Keyword	Range
`Int8`	–128 to 127
`Int16`	–32,768 to 32,767
`Int32`	–2,147,483,648 to 2,147,483,647
`Int64`	–9,223,372,036,854,775,808 to 9,223,372,036,854,775,807

For unsigned integers (UInt8, UInt16, UInt32, UInt64), see Integer types in the ClickHouse documentation.

Floating-point

Keyword	Description
`Float32`	Single-precision floating-point. Equivalent to `FLOAT` in C++. Takes 4 bytes; 32-bit binary.
`Float64`	Double-precision floating-point. Equivalent to `DOUBLE` in C++. Takes 8 bytes; 64-bit binary.

For financial or business data that requires exact calculations, use Decimal instead of Float32 or Float64. Floating-point arithmetic can introduce rounding errors.

Decimal

Signed fixed-point numbers that maintain precision through addition, subtraction, and multiplication.

Supported formats: Decimal(P, S), Decimal32(S), Decimal64(S), Decimal128(S).

String

Keyword	Description
`String`	Variable-length string of any size, including null bytes. Equivalent to `VARCHAR`, `BLOB`, and `CLOB` in other database systems.
`FixedString(N)`	Fixed-length string of exactly N bytes. More efficient than `String` when all values have the same byte length.

Efficient use cases for FixedString:

Binary representation of IPv6 addresses: FixedString(16)
Language codes such as ru_RU and en_US
Currency codes such as USD and RUB
MD5 hash values: FixedString(16)
SHA-256 hash values: FixedString(32)

Date and time

Keyword	Description
`Date`	Stores a date as the number of days since 1970-01-01. Takes 2 bytes. Unsigned. Stored without a time zone.
`DateTime`	Stores a datetime as the number of seconds since 1970-01-01 00:00:00. Takes 4 bytes. Unsigned. The value range of `DateTime` is the same as the value range of `Date`. The minimum value is 1970-01-01 00:00:00. Accurate to seconds; leap seconds are not supported. The time zone is the system time zone of the client or server.
`DateTime64`	Stores both the date and time of a specific point in time.

Date is faster than DateTime under most conditions and requires half the storage (2 bytes vs. 4 bytes). The size difference becomes more significant after compression.

Boolean

ClickHouse does not have a native Boolean type. Use UInt8 for Boolean values: 0 for false, 1 for true.

Array

Array(T) — a collection of elements of the same type T. The element type can be any ClickHouse type, including Array.

Avoid multi-dimensional arrays. ClickHouse supports them only partially — multi-dimensional arrays cannot be stored in MergeTree tables.

Tuple

Tuple(T1, T2, ...) — a collection of elements of different types.

Tuples cannot be stored in tables except in memory tables. Use them to group temporary columns in queries with IN expressions or lambda functions.

Domain

The Domain type stores IP addresses in binary format with human-readable input and output:

IPv4 addresses: binary-compatible with UInt32.
IPv6 addresses: binary-compatible with FixedString(16).

Enumeration

Keyword	Range	Notes
`Enum8`	–128 to 127	—
`Enum16`	–32,768 to 32,767	Values can be empty

Nullable

Nullable(T) — wraps any base type T to allow NULL values. The default value is NULL unless the ClickHouse server configuration specifies otherwise.

Constraints:

Nullable columns cannot be included in table indexes.

Nested

Nested — a nested data structure, similar to a table embedded in a cell. Define it with field names and types, using the same syntax as CREATE TABLE. Each row in the outer table can correspond to any number of rows in the nested structure.