HCL Column Types

The following guide describes the column types supported by Atlas HCL, and how to use them.

MySQL

Bit

The bit type allows creating BIT columns. An optional size attribute allows controlling the number of bits stored in a column, ranging from 1 to 64.

table "t" {
  schema = schema.test
  column "c1" {
    type = bit
  }
  column "c2" {
    type = bit(4)
  }
}

Binary

The varbinary and binary types allow storing binary byte strings.

table "t" {
  schema = schema.test
  column "c1" {
    // Equals to binary(1).
    type = binary
  }
  column "c2" {
    type = binary(10)
  }
  column "c3" {
    type = varbinary(255)
  }
}

Blob

The tinyblob, mediumblob, blob and longblob types allow storing binary large objects.

table "t" {
  schema = schema.test
  column "c1" {
    type = tinyblob
  }
  column "c2" {
    type = mediumblob
  }
  column "c3" {
    type = blob
  }
  column "c4" {
    type = longblob
  }
}

Boolean

The bool and boolean types are mapped to tinyint(1) in MySQL. Still, Atlas allows maintaining columns of type bool in the schema for simplicity reasons.

table "t" {
  schema = schema.test
  column "c1" {
    type = bool
  }
  column "c2" {
    type = boolean
  }
}

Learn more about the motivation for these types in the MySQL website.

Date and Time

Atlas supports the standard MySQL types for storing date and time values: time, timestamp, date, datetime, and year.

table "t" {
  schema = schema.test
  column "c1" {
    type = time
  }
  column "c2" {
    type = timestamp
  }
  column "c3" {
    type = date
  }
  column "c4" {
    type = datetime
  }
  column "c5" {
    type = year
  }
  column "c6" {
    type = time(1)
  }
  column "c7" {
    type = timestamp(2)
  }
  column "c8" {
    type = datetime(4)
  }
}

Fixed Point (Decimal)

The decimal and numeric types are supported for storing exact numeric values. Note that in MySQL the two types are identical.

table "t" {
  schema = schema.test
  column "c1" {
    // Equals to decimal(10) as the
    // default precision is 10.
    type = decimal
  }
  column "c2" {
    // Equals to decimal(5,0).
    type = decimal(5)
  }
  column "c3" {
    type = decimal(5,2)
  }
  column "c4" {
    type     = numeric
    unsigned = true
  }
}

Floating Point (Float)

The float and double types are supported for storing approximate numeric values.

table "t" {
  schema = schema.test
  column "c1" {
    type = float
  }
  column "c2" {
    type = double
  }
  column "c3" {
    type     = float
    unsigned = true
  }
  column "c4" {
    type     = double
    unsigned = true
  }
}

Enum

The enum type allows storing a set of enumerated values.

table "t" {
  schema = schema.test
  column "c1" {
    type = enum("a", "b")
  }
  column "c2" {
    type = enum(
      "c",
      "d",
    )
  }
}

Integer

The tinyint, smallint, int, mediumint, bigint integer types are support by Atlas.

table "t" {
  schema = schema.test
  column "c1" {
    type = int
  }
  column "c2" {
    type = tinyint
  }
  column "c3" {
    type = smallint
  }
  column "c4" {
    type = mediumint
  }
  column "c5" {
    type = bigint
  }
}

Integer Attributes

The auto_increment, and unsigned attributes are also supported by integer types.

table "t" {
  schema = schema.test
  column "c1" {
    type     = tinyint
    unsigned = true
  }
  column "c2" {
    type           = smallint
    auto_increment = true
  }
  primary_key {
    columns = [column.c2]
  }
}

JSON

The json type allows storing JSON objects.

table "t" {
  schema = schema.test
  column "c1" {
    type = json
  }
}

Set

The set type allows storing a set of values.

table "t" {
  schema = schema.test
  column "c1" {
    type = set("a", "b")
  }
  column "c2" {
    type = set(
      "c",
      "d",
    )
  }
}

String

Atlas supports the standard MySQL types for storing string values. varchar, char, tinytext, mediumtext, text and longtext.

table "t" {
  schema = schema.test
  column "c1" {
    type = varchar(255)
  }
  column "c2" {
    type = char(1)
  }
  column "c3" {
    type = tinytext
  }
  column "c4" {
    type = mediumtext
  }
  column "c5" {
    type = text
  }
  column "c6" {
    type = longtext
  }
}

Spatial

The geometry, point, multipoint, linestring and the rest of the MySQL spatial types are supported by Atlas.

table "t" {
  schema = schema.test
  column "c1" {
    type = geometry
  }
  column "c2" {
    type = point
  }
  column "c3" {
    type = multipoint
  }
  column "c4" {
    type = linestring
  }
}

PostgreSQL

Array

Atlas supports defining PostgreSQL array types using the sql function.

table "t" {
  schema = schema.test
  column "c1" {
    type = sql("int[]")
  }
  column "c2" {
    type = sql("text[]")
  }
  column "c3" {
    type = sql("int ARRAY")
  }
  column "c4" {
    type = sql("varchar(255)[]")
  }
  column "c5" {
    // The current PostgreSQL implementation
    // ignores any supplied array size limits.
    type = sql("point[4][4]")
  }
}

Bit

The bit and bit varying types allow creating bit string columns.

table "t" {
  schema = schema.test
  column "c1" {
    // Equals to bit(1).
    type = bit
  }
  column "c2" {
    type = bit(2)
  }
  column "c3" {
    // Unlimited length.
    type = bit_varying
  }
  column "c4" {
    type = bit_varying(1)
  }
}

Boolean

The boolean type allows creating standard SQL boolean columns.

table "t" {
  schema = schema.test
  column "c1" {
    type = boolean
  }
  column "c2" {
    type    = boolean
    default = true
  }
}

Binary

The bytea type allows creating binary string columns.

table "t" {
  schema = schema.test
  column "c1" {
    type = bytea
  }
}

Date, Time and Interval

Atlas supports the standard PostgreSQL types for creating date, time and interval columns.

table "t" {
  schema = schema.test
  column "c1" {
    type = date
  }
  column "c2" {
    // Equals to "time without time zone".
    type = time
  }
  column "c3" {
    // Equals to "time with time zone".
    type = timetz
  }
  column "c4" {
    // Equals "timestamp without time zone".
    type = timestamp
  }
  column "c5" {
    // Equals "timestamp with time zone".
    type = timestamptz
  }
  column "c6" {
    type = timestamp(4)
  }
  column "c7" {
    type = interval
  }
}

Domain

The domain type is a user-defined data type that is based on an existing data type but with optional constraints and default values. Learn more about it in the PostgreSQL website.

domain "us_postal_code" {
  schema = schema.public
  type   = text
  null   = true
  check "us_postal_code_check" {
    expr = "((VALUE ~ '^\\d{5}$'::text) OR (VALUE ~ '^\\d{5}-\\d{4}$'::text))"
  }
}

domain "username" {
  schema = schema.public
  type    = text
  null    = false
  default = "anonymous"
  check "username_length" {
    expr = "(length(VALUE) > 3)"
  }
}

table "users" {
  schema = schema.public
  column "name" {
    type = domain.username
  }
  column "zip" {
    type = domain.us_postal_code
  }
}

Enum

The enum type allows storing a set of enumerated values. Learn more about it in the PostgreSQL website.

enum "status" {
  schema = schema.test
  values = ["on", "off"]
}

table "t1" {
  schema = schema.test
  column "c1" {
    type = enum.status
  }
}

table "t2" {
  schema = schema.test
  column "c1" {
    type = enum.status
  }
}

Fixed Point (Decimal)

The decimal and numeric types are supported for storing exact numeric values. Note that in PostgreSQL the two types are identical.

table "t" {
  schema = schema.test
  column "c1" {
    // Equals to decimal.
    type = numeric
  }
  column "c2" {
    // Equals to decimal(5).
    type = numeric(5)
  }
  column "c3" {
    // Equals to decimal(5,2).
    type = numeric(5,2)
  }
}

Floating Point (Float)

The real and double_precision types are supported for storing approximate numeric values.

table "t" {
  schema = schema.test
  column "c1" {
    type = real
  }
  column "c2" {
    type = double_precision
  }
  column "c3" {
    // Equals to real when precision is between 1 to 24.
    type = float(10)
  }
  column "c2" {
    // Equals to double_precision when precision is between 1 to 24.
    type = float(30)
  }
}

Geometric

Atlas supports the standard PostgreSQL types for creating geometric columns.

table "t" {
  schema = schema.test
  column "c1" {
    type = circle
  }
  column "c2" {
    type = line
  }
  column "c3" {
    type = lseg
  }
  column "c4" {
      type = box
  }
  column "c5" {
      type = path
  }
  column "c6" {
      type = polygon
  }
  column "c7" {
      type = point
  }
}

Integer

The smallint, integer / int, bigint types allow creating integer types.

table "t" {
  schema = schema.test
  column "c1" {
    type = smallint
  }
  column "c2" {
    type = integer
  }
  column "c3" {
    type = int
  }
  column "c4" {
    type    = bigint
    default = 1
  }
}

JSON

The json and jsonb types allow creating columns for storing JSON objects.

table "t" {
  schema = schema.test
  column "c1" {
    type = json
  }
  column "c2" {
    type = jsonb
  }
}

Money

The money data type allows creating columns for storing currency amount with a fixed fractional precision.

table "t" {
  schema = schema.test
  column "c1" {
    type = money
  }
}

Network Address

The inet, cidr, macaddr and macaddr8 types allow creating network address columns.

table "t" {
  schema = schema.test
  column "c1" {
    type = inet
  }
  column "c2" {
    type = cidr
  }
  column "c3" {
    type = macaddr
  }
  column "c4" {
    type = macaddr8
  }
}

Range

PostgreSQL supports the creation of range types for storing range of values of some element type. Learn more about them in the PostgreSQL website.

table "t" {
  schema = schema.test
  column "r1" {
    type = int4range
  }
  column "r2" {
    type = int8range
  }
  column "r3" {
    type = numrange
  }
  column "r4" {
    type = tsrange
  }
  column "r5" {
    type = tstzrange
  }
  column "r6" {
    type = daterange
  }
  column "r7" {
    type = int4multirange
  }
  column "r8" {
    type = int8multirange
  }
  column "r9" {
    type = nummultirange
  }
  column "r10" {
    type = tsmultirange
  }
  column "r11" {
    type = tstzmultirange
  }
  column "r12" {
    type = datemultirange
  }
}

Serial

PostgreSQL supports creating columns of types smallserial, serial, and bigserial. Note that these types are not actual types, but more like "macros" for creating non-nullable integer columns with sequences attached.

table "t" {
  schema = schema.test
  column "c1" {
      type = smallserial
  }
  column "c2" {
      type = serial
  }
  column "c3" {
      type = bigserial
  }
}

String

The varchar, char, character_varying, character and text types allow creating string columns.

table "t" {
  schema = schema.test
  column "c1" {
    // Unlimited length.
    type = varchar
  }
  column "c2" {
    // Alias to character_varying(255).
    type = varchar(255)
  }
  column "c3" {
    // Equals to char(1).
    type = char
  }
  column "c4" {
    // Alias to character(5).
    type = char(5)
  }
  column "c5" {
    type = text
  }
}

Text Search

The tsvector and tsquery data types are designed to store and query full text search. Learn more about them in the PostgreSQL website.

table "t" {
  schema = schema.test
  column "tsv" {
    type = tsvector
  }
  column "tsq" {
    type = tsquery
  }
}

UUID

The uuid data type allows creating columns for storing Universally Unique Identifiers (UUID).

table "t" {
  schema = schema.test
  column "c1" {
    type = uuid
  }
  column "c2" {
    type    = uuid
    default = sql("gen_random_uuid()")
  }
}

XML

The xml data type allows creating columns for storing XML data.

table "t" {
  schema = schema.test
  column "c1" {
    type = xml
  }
}

SQLite

Values in SQLite are stored in one of the four native types: BLOB, INTEGER, NULL, TEXT and REAL. Still, Atlas supports variety of data types that are commonly used by ORMs. These types are mapped to column affinities based on the rules described in SQLite website.

Blob

The blob data type allows creating columns with BLOB type affinity.

table "t" {
  schema = schema.main
  column "c" {
    type = blob
  }
}

Integer

The int and integer data types allow creating columns with INTEGER type affinity.

table "t" {
  schema = schema.main
  column "c" {
    type = int
  }
}

Numeric

The numeric and decimal data types allow creating columns with NUMERIC type affinity.

table "t" {
  schema = schema.main
  column "c" {
    type = decimal
  }
}

Text

The text, varchar, clob, character and varying_character data types allow creating columns with text type affinity. i.e. stored as text strings.

table "t" {
  schema = schema.main
  column "c" {
    type = text
  }
}

Real

The real, double, double_precision, and float data types allow creating columns with real type affinity.

table "t" {
  schema = schema.main
  column "c" {
    type = real
  }
}

Additional Types

As mentioned above, Atlas supports variety of data types that are commonly used by ORMs. e.g. Ent.

table "t" {
  schema = schema.main
  column "c1" {
    type = bool
  }
  column "c2" {
    type = date
  }
  column "c3" {
    type = datetime
  }
  column "c4" {
    type = uuid
  }
  column "c5" {
    type = json
  }
}

SQL Server

Bit

The bit type allows creating BIT columns.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = bit
  }
}

Binary strings

The varbinary and binary types allow storing binary byte strings.

table "t" {
  schema = schema.dbo
  column "c1" {
    // Equals to binary(1).
    type = binary
  }
  column "c2" {
    type = binary(10)
  }
  column "c3" {
    type = varbinary(255)
  }
  column "c4" {
    // Max length: 8,000 bytes.
    type = varbinary(MAX)
  }
}

Date and Time

Atlas supports the standard SQL Server types for storing date and time values: date, datetime, datetime2, datetimeoffset, smalldatetime and time. The document on Microsoft website has more information on date and time types.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = date
  }
  column "c2" {
    type = datetime
  }
  column "c3" {
    type = datetime2
  }
  column "c4" {
    type = datetimeoffset
  }
  column "c5" {
    type = smalldatetime
  }
  column "c6" {
    // Equals to time(7).
    type = time
  }
  column "c7" {
    type = time(1)
  }
  column "c8" {
    type = time(2)
  }
  column "c9" {
    type = time(3)
  }
  column "c10" {
    type = time(4)
  }
  column "c11" {
    type = time(5)
  }
  column "c12" {
    type = time(6)
  }
}

Integer

The int, bigint, smallint, and tinyint integer types are support by Atlas. See document on Microsoft website for more information on integer types.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = int
  }
  column "c2" {
    type = tinyint
  }
  column "c3" {
    type = smallint
  }
  column "c4" {
    type = bigint
  }
}

Integer Blocks

The identity block can be used to create an identity column.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = tinyint
  }
  column "c2" {
    type = bigint
    identity {
      seed      = 701
      increment = 1000
    }
  }
  primary_key {
    columns = [column.c2]
  }
}

Fixed Point (Decimal)

The decimal and numeric types are supported for storing exact numeric values. Note that in SQL Server the two types are identical.

table "t" {
  schema = schema.dbo
  column "c1" {
    // Equals to decimal(18, 0) as the
    // default precision is 18.
    type = decimal
  }
  column "c2" {
    // Equals to decimal(5,0).
    type = decimal(5)
  }
  column "c3" {
    type = decimal(5,2)
  }
  column "c4" {
    type = numeric
  }
}

Floating Point (Float)

The float and real types are supported for storing approximate numeric values. The document on Microsoft website has more information on float types.

table "t" {
  schema = schema.dbo
  column "c1" {
    // Equals to float(53).
    type = float
  }
  column "c2" {
    // float(n) is between 1 and 53.
    type = float(12)
  }
  column "c3" {
    // The ISO synonym for real is `float(24)`.
    type = real
  }
}

Money

The money and smallmoney data types allows creating columns for storing currency amount with a fixed fractional precision.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = money
  }
  column "c2" {
    type = smallmoney
  }
}

Character strings

The char, and varchar types allow creating string columns. The document on Microsoft website has more information on string types.

table "t" {
  schema = schema.dbo
  column "c1" {
    // Equals to varchar(1).
    type = varchar
  }
  column "c2" {
    type = varchar(255)
  }
  column "c3" {
    type = varchar(MAX)
  }
  column "c4" {
    // Equals to char(1).
    type = char
  }
  column "c5" {
    type = char(5)
  }
}

Unicode character strings

The nchar, and nvarchar types allow creating string columns. The document on Microsoft website has more information on unicode string types.

table "t" {
  schema = schema.dbo
  column "c1" {
    // Equals to nvarchar(1).
    type = nvarchar
  }
  column "c2" {
    type = nvarchar(255)
  }
  column "c3" {
    type = nvarchar(MAX)
  }
  column "c4" {
    // Equals to nchar(1).
    type = nchar
  }
  column "c5" {
    type = nchar(5)
  }
}

`ntext`, `text` and `image`

Atlas supports some deprecated types for backward compatibility. The document on Microsoft website has more information on ntext, text and image types.

table "t" {
  schema = schema.dbo
  column "c1" {
    type = ntext
  }
  column "c2" {
    type = text
  }
  column "c3" {
    type = image
  }
}

User-defined types

There are two types of user-defined types are supported by Atlas: Alias Types and Table Types.

The CLR user-defined types are not supported by Atlas.

Alias Types

The type_alias type allows creating columns with user-defined types.

type_alias "ssn" {
  schema = schema.dbo
  type   = varchar(11)
  null   = false
}
type_alias "age" {
  schema = schema.dbo
  type   = smallint
  null   = false
}
table "t" {
  schema = schema.dbo
  column "ssn" {
    type = type_alias.ssn
  }
  column "age" {
    type = type_alias.age
  }
}

Table Types

The type_table type allows the creation of columns with user-defined table types. The User-Defined table type only allows to use of functions/procedures arguments and not on table columns.

type_table "address" {
  schema = schema.dbo
  column "ssn" {
    type = type_alias.ssn
  }
  column "street" {
    type = varchar(255)
  }
  column "city" {
    type = varchar(255)
  }
  column "state" {
    type = varchar(2)
  }
  column "zip" {
    type = type_alias.zip
  }
  index {
    unique  = true
    columns = [column.ssn]
  }
  check "zip_check" {
    expr = "len(zip) = 5"
  }
}
function "fn1" {
  schema = schema.dbo
  lang   = SQL
  arg "@a1" {
    type     = type_table.address
    readonly = true // The table type is readonly argument.
  }
  arg "@zip" {
    type = type_alias.zip
  }
  return = int
  as     = <<-SQL
  BEGIN
    RETURN (SELECT COUNT(1) FROM @a1 WHERE zip = @zip);
  END
  SQL
}
type_alias "ssn" {
  schema = schema.foo
  type   = varchar(11)
  null   = false
}
type_alias "zip" {
  schema = schema.foo
  type   = varchar(5)
  null   = false
}

note

SQL Server doesn't support creating a named unique constraint on a user-defined table type. Atlas was unable to handle duplicate unique constraints (the unique constraints on the same columns) on table types. The below example will cause schema diff for every time it applies schema.

CREATE TYPE [typ1] AS TABLE (
  [c1] int NOT NULL UNIQUE ([c1] DESC),
  UNIQUE ([c1] ASC)
);

ClickHouse

Array

Atlas supports defining ClickHouse array types using the sql function.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    type = sql("Array(Int32)")
  }
  column "c2" {
    type = sql("Array(String)")
  }
  column "c3" {
    type = sql("Array(Array(Int32))")
  }
}

Boolean

The Bool type allows creating standard SQL boolean columns.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    type = Bool
  }
  column "c2" {
    type    = Bool
    default = true
  }
}

Date and Time

Atlas supports the standard ClickHouse types for creating date and time columns: Date, DateTime, DateTime32 DateTime64.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null     = false
    type     = Date
  }
  column "c2" {
    null    = false
    type    = Date32
  }
  column "c3" {
    null    = false
    type    = DateTime
  }
  column "c4" {
    null    = false
    type    = DateTime("America/New_York")
  }
  column "c5" {
    null    = false
    type    = DateTime
  }
  column "c6" {
    null    = false
    type    = DateTime32("America/New_York")
  }
  column "c7" {
    null    = false
    type    = DateTime64(3)
  }
  column "c8" {
    null    = false
    type    = DateTime64(3, "America/New_York")
  }
}

Fixed Point (Decimal)

The Decimal type allows creating columns for storing exact numeric values. The precision and scale are specified as below.

Decimal Precision: 9, Scale: 0
Decimal32(Scale) Precision: 9, Scale: Scale
Decimal64(Scale) Precision: 18, Scale: Scale
Decimal128(Scale) Precision: 38, Scale: Scale
Decimal256(Scale) Precision: 76, Scale: Scale
Decimal(Precision, Scale) Precision: Precision, Scale: Scale

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Decimal
  }
  column "c2" {
    null = false
    type = Decimal32(2)
  }
  column "c3" {
    null = false
    type = Decimal64(2)
  }
  column "c4" {
    null = false
    type = Decimal128(2)
  }
  column "c5" {
    null = false
    type = Decimal256(2)
  }
  column "c6" {
    null = false
    type = Decimal(11, 2)
  }
}

Enum

The Enum type allows storing a set of enumerated values and supports defining ClickHouse enum types using the sql function.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Enum("a", "b")
  }
  column "c2" {
    null = false
    type = Enum8("a", "b")
  }
  column "c3" {
    null = false
    type = Enum16("a", "b")
  }
}

Fixed String

The FixedString type allows creating columns for storing fixed-length string values.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = FixedString(10)
  }
}

Floating Point (Float)

The Float32 and Float64 types are supported for storing approximate numeric values. The aliases for these types are Float and Double.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Float
  }
  column "c2" {
    null = false
    type = Double
  }
}

Integer

The Int8, Int16, Int32, Int64, Int128, Int256 types allow creating integer types. The aliases for these types are Tinyint, Smallint, Int, Bigint.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Tinyint
  }
  column "c2" {
    null = false
    type = Smallint
  }
  column "c3" {
    null = false
    type = Int
  }
  column "c4" {
    null = false
    type = Bigint
  }
  column "c5" {
    null = false
    type = Int128
  }
  column "c6" {
    null = false
    type = Int256
  }
}

Integer Attributes

The Unsigned attribute is also supported by integer types.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null     = false
    type     = Int
    unsigned = true
  }
}

IPv4 and IPv6

The IPv4 and IPv6 types allow creating columns for storing IPv4 and IPv6 addresses.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = IPv4
  }
  column "c2" {
    null = false
    type = IPv6
  }
}

Spatial

Atlas supports the standard ClickHouse types for creating spatial columns.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Point
  }
  column "c2" {
    null = false
    type = Polygon
  }
  column "c3" {
    null = false
    type = MultiPolygon
  }
}

Ring

The Ring type allows creating columns for storing ring values.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = Ring
  }
}

String

The String type allows creating columns for storing string values.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = String
  }
}

UUID

The UUID type allows creating columns for storing Universally Unique Identifiers (UUID).

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = UUID
  }
}

Tuple

Atlas supports defining ClickHouse tuple types using the sql function.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = sql("Tuple(Int32, String)")
  }
}

LowCardinality

Atlas supports defining ClickHouse low cardinality types using the sql function.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = sql("LowCardinality(String)")
  }
}

Nullable

Atlas supports defining ClickHouse nullable types using the sql function. Null attribute is needed to be set to true for nullable types.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = true
    type = sql("Nullable(String)")
  }
}

JSON

The JSON type allows creating columns for storing JSON objects.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    type = JSON
  }
}

AggregateFunction

Atlas supports defining ClickHouse aggregate data types by using the sql function.

table "t" {
  schema = schema.test
  engine = Memory
  column "c1" {
    null = false
    type = sql("AggregateFunction(uniq, UInt64)")
  }
  column "c2" {
    null = false
    type = sql("SimpleAggregateFunction(sum, Int32)")
  }
}

info

The AggregateFunction and SimpleAggregateFunction are complex data types. Therefore, we recommend using a Dev Database to normalize these types.

MySQL​

Bit​

Binary​

Blob​

Boolean​

Date and Time​

Fixed Point (Decimal)​

Floating Point (Float)​

Enum​

Integer​

Integer Attributes​

JSON​

Set​

String​

Spatial​

PostgreSQL​

Array​

Bit​

Boolean​

Binary​

Date, Time and Interval​

Domain​

Enum​

Fixed Point (Decimal)​

Floating Point (Float)​

Geometric​

Integer​

JSON​

Money​

Network Address​

Range​

Serial​

String​

Text Search​

UUID​

XML​

SQLite​

Blob​

Integer​

Numeric​

Text​

Real​

Additional Types​

SQL Server​

Bit​

Binary strings​

Date and Time​

Integer​

Integer Blocks​

Fixed Point (Decimal)​

Floating Point (Float)​

Money​

Character strings​

Unicode character strings​

ntext, text and image​

User-defined types​

Alias Types​

Table Types​

ClickHouse​

Array​

Boolean​

Date and Time​

Fixed Point (Decimal)​

Enum​

Fixed String​

Floating Point (Float)​

Integer​

Integer Attributes​

IPv4 and IPv6​

Spatial​

Ring​

String​

UUID​

Tuple​

LowCardinality​

Nullable​

JSON​

AggregateFunction​

MySQL

Bit

Binary

Blob

Boolean

Date and Time

Fixed Point (Decimal)

Floating Point (Float)

Enum

Integer

Integer Attributes

JSON

Set

String

Spatial

PostgreSQL

Array

Bit

Boolean

Binary

Date, Time and Interval

Domain

Enum

Fixed Point (Decimal)

Floating Point (Float)

Geometric

Integer

JSON

Money

Network Address

Range

Serial

String

Text Search

UUID

XML

SQLite

Blob

Integer

Numeric

Text

Real

Additional Types

SQL Server

Bit

Binary strings

Date and Time

Integer

Integer Blocks

Fixed Point (Decimal)

Floating Point (Float)

Money

Character strings

Unicode character strings

`ntext`, `text` and `image`

User-defined types

Alias Types

Table Types

ClickHouse

Array

Boolean

Date and Time

Fixed Point (Decimal)

Enum

Fixed String

Floating Point (Float)

Integer

Integer Attributes

IPv4 and IPv6

Spatial

Ring

String

UUID

Tuple

LowCardinality

Nullable

JSON

AggregateFunction