SQL Column Types

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
  }
}

Enum

The enum type allows storing a set of enumerated values.

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
  }
}