CREATE_TYPE(7pg16)

📦 version 16.15, in opensuse-leap 16.0

💅 CREATE TYPE 💅

CREATE TYPE — a.k.a. the PostgreSQL 16.15 documentation glow-up for define a new data type. Yes bestie, this is the manual page, and we are making it make sense.

✨ NAME ✨

CREATE_TYPE — define a new data type 🎀

✨ SYNOPSIS ✨

Okay so here's the whole vibe check, all five forms, copied exactly because I would never mess with your syntax:

CREATE TYPE name AS
    ( [ attribute_name data_type [ COLLATE collation ] [, ... ] ] )

CREATE TYPE name AS ENUM
    ( [ 'label' [, ... ] ] )

CREATE TYPE name AS RANGE (
    SUBTYPE = subtype
    [ , SUBTYPE_OPCLASS = subtype_operator_class ]
    [ , COLLATION = collation ]
    [ , CANONICAL = canonical_function ]
    [ , SUBTYPE_DIFF = subtype_diff_function ]
    [ , MULTIRANGE_TYPE_NAME = multirange_type_name ]
)

CREATE TYPE name (
    INPUT = input_function,
    OUTPUT = output_function
    [ , RECEIVE = receive_function ]
    [ , SEND = send_function ]
    [ , TYPMOD_IN = type_modifier_input_function ]
    [ , TYPMOD_OUT = type_modifier_output_function ]
    [ , ANALYZE = analyze_function ]
    [ , SUBSCRIPT = subscript_function ]
    [ , INTERNALLENGTH = { internallength | VARIABLE } ]
    [ , PASSEDBYVALUE ]
    [ , ALIGNMENT = alignment ]
    [ , STORAGE = storage ]
    [ , LIKE = like_type ]
    [ , CATEGORY = category ]
    [ , PREFERRED = preferred ]
    [ , DEFAULT = default ]
    [ , ELEMENT = element ]
    [ , DELIMITER = delimiter ]
    [ , COLLATABLE = collatable ]
)

CREATE TYPE name

🌸 DESCRIPTION 🌸

So CREATE TYPE registers a brand-new data type for use in the current database, and the user who defines the type becomes its owner. Main character behavior honestly 👑.

If you give a schema name, the type gets created in that schema. If you don't, it lands in the current schema. And the type name has to be distinct from the name of any existing type or domain in the same schema. (Also — because tables have associated data types, the type name must also be distinct from the name of any existing table in the same schema. So no name-stealing, bestie 🚫.)

There are five forms of CREATE TYPE, exactly like the syntax synopsis up there shows. In order, they create:

  1. a composite type
  2. an enum type
  3. a range type
  4. a base type
  5. a shell type

The first four get discussed one by one below 🫶. A shell type is literally just a placeholder for a type you'll define later — you make one by issuing CREATE TYPE with no parameters except the type name. Shell types are what you need for forward references when creating range types and base types, and we get into that in those sections, iykyk 🐛.

🦋 Composite Types 🦋

The first form of CREATE TYPE creates a composite type. You spec it out with a list of attribute names and data types. An attribute's collation can be specified too, if its data type is collatable. A composite type is essentially the same thing as the row type of a table — but using CREATE TYPE means you skip creating an actual table when all you wanted was to define a type. A stand-alone composite type is genuinely useful, like, for example, as the argument or return type of a function ✨.

To be able to create a composite type, you must have USAGE privilege on all attribute types. All of them. No cap.

🎀 Enumerated Types 🎀

The second form of CREATE TYPE creates an enumerated (enum) type, as described in Section 8.7. Enum types take a list of quoted labels, and each one must be less than NAMEDATALEN bytes long (64 bytes in a standard PostgreSQL build). And yes — it is possible to create an enumerated type with zero labels, but such a type cannot be used to hold values before at least one label is added using ALTER TYPE. So like... technically valid, practically a vibe-less empty box 🙈.

💜 Range Types 💜

The third form of CREATE TYPE creates a new range type, as described in Section 8.17.

The range type's subtype can be any type with an associated b-tree operator class — that's what determines the ordering of values for the range type. Normally the subtype's default b-tree operator class handles the ordering; if you want a non-default operator class, specify its name with subtype_opclass. And if the subtype is collatable and you want a non-default collation in the range's ordering, name the collation you want with the collation option 🎀.

The optional canonical function must take one argument of the range type being defined, and return a value of that same type. It converts range values to a canonical form when applicable. Peep Section 8.17.8 for more. Real talk: creating a canonical function is a bit tricky, because it has to be defined before the range type can be declared. The way you pull that off: first create a shell type, which is a placeholder type with no properties except a name and an owner. You do that by issuing the command CREATE TYPE name with no additional parameters. Then you can declare the function using the shell type as argument and result, and finally declare the range type using the same name. That automatically replaces the shell type entry with a valid range type. Iconic little switcheroo 🔥.

The optional subtype_diff function must take two values of the subtype type as arguments, and return a double precision value representing the difference between the two given values. It's optional, but supplying it allows much greater efficiency of GiST indexes on columns of the range type. Section 8.17.8 again, bestie 📖.

The optional multirange_type_name parameter specifies the name of the corresponding multirange type. If you don't specify it, the name gets chosen automatically like this:

  • If the range type name contains the substring range, then the multirange type name is formed by replacement of that range substring with multirange in the range type name.
  • Otherwise, the multirange type name is formed by appending a _multirange suffix to the range type name.
To be able to create a range type, you must have USAGE privilege on the subtype.

🧚 Base Types 🧚

The fourth form of CREATE TYPE creates a new base type (scalar type). And here's the thing — to create a new base type, you must be a superuser. 💀 (This restriction exists because an erroneous type definition could confuse or even crash the server. Not a drill, bestie, this is the dangerous one.)

The parameters can appear in any order, not only the order illustrated above, and most of them are optional. You must register two or more functions (using CREATE FUNCTION) before defining the type. The support functions input_function and output_function are required, while receive_function, send_function, type_modifier_input_function, type_modifier_output_function, analyze_function, and subscript_function are optional. Generally these functions have to be coded in C or another low-level language 🫠.

The input_function converts the type's external textual representation to the internal representation used by the operators and functions defined for the type. The output_function performs the reverse transformation. The input function can be declared as taking one argument of type cstring, or as taking three arguments of types cstring, oid, integer. Let's break those three down:

  • the first argument is the input text as a C string
  • the second argument is the type's own OID (except for array types, which instead receive their element type's OID)
  • the third is the typmod of the destination column, if known (-1 will be passed if not)

The input function must return a value of the data type itself. Usually an input function should be declared STRICT; if it is not, it will be called with a NULL first parameter when reading a NULL input value. The function must still return NULL in this case, unless it raises an error. (This case is mainly meant to support domain input functions, which might need to reject NULL inputs.) The output function must be declared as taking one argument of the new data type. The output function must return type cstring. Output functions are not invoked for NULL values.

The optional receive_function converts the type's external binary representation to the internal representation. If you don't supply this function, the type cannot participate in binary input. Your binary representation should be chosen to be cheap to convert to internal form, while still being reasonably portable. (For example, the standard integer data types use network byte order as the external binary representation, while the internal representation is in the machine's native byte order. Cute, right? 🦋) The receive function should perform adequate checking to make sure the value is valid. It can be declared as taking one argument of type internal, or as taking three arguments of types internal, oid, integer. The first argument is a pointer to a StringInfo buffer holding the received byte string; the optional arguments are the same as for the text input function. The receive function must return a value of the data type itself. Usually a receive function should be declared STRICT; if it is not, it will be called with a NULL first parameter when reading a NULL input value. The function must still return NULL in this case, unless it raises an error. (This case is mainly meant to support domain receive functions, which might need to reject NULL inputs.)

Similarly, the optional send_function converts from the internal representation to the external binary representation. If this function isn't supplied, the type cannot participate in binary output. The send function must be declared as taking one argument of the new data type. The send function must return type bytea. Send functions are not invoked for NULL values.

You should at this point be wondering: how on earth can the input and output functions be declared to have results or arguments of the new type, when they have to be created before the new type can be created?? 🤔 The answer, bestie, is that the type should first be defined as a shell type — that placeholder type with no properties except a name and an owner, remember? You make it by issuing the command CREATE TYPE name with no additional parameters. Then the C I/O functions can be defined referencing the shell type. And finally, CREATE TYPE with a full definition replaces the shell entry with a complete, valid type definition, after which the new type can be used normally. Full circle, ate and left no crumbs ✨🍓.

The optional type_modifier_input_function and type_modifier_output_function are needed if your type supports modifiers — those optional constraints you tack onto a type declaration, like char(5) or numeric(30,2). ✨ PostgreSQL lets user-defined types take one or more simple constants or identifiers as modifiers, but bestie, here's the catch: that info has to be packable into a single non-negative integer value so it can live in the system catalogs.

So the type_modifier_input_function gets handed the declared modifier(s) as a cstring array. It must check the values for validity (throw an error if they're wrong 😤), and if they're correct, return a single non-negative integer that gets stored as the column "typmod". Note that type modifiers will be rejected outright if the type doesn't have a type_modifier_input_function.

The type_modifier_output_function goes the other way — it converts the internal integer typmod back into the right form for user display. It must return a cstring value that is the exact string to append to the type name; for example numeric's function might return (30,2). It's totally allowed to omit the output function, in which case the default display format is just the stored typmod integer value wrapped in parentheses. 💅

The optional analyze_function performs type-specific statistics collection for columns of the data type. By default, ANALYZE will attempt to gather statistics using the type's "equals" and "less-than" operators, if there's a default b-tree operator class for the type. For non-scalar types this behavior is likely to be unsuitable 🫠, so it can be overridden by specifying a custom analysis function. The analysis function has to be declared to take a single argument of type internal, and return a boolean result. The detailed API for analysis functions appears in src/include/commands/vacuum.h.

The optional subscript_function lets the data type be subscripted in SQL commands. 🌈 Specifying this function does not make the type a "true" array type; for example, it won't be a candidate for the result type of ARRAY[] constructs. But if subscripting a value of the type is a natural notation for extracting data from it, then a subscript_function can be written to define what that means. The subscript function must be declared to take a single argument of type internal, and return an internal result, which is a pointer to a struct of methods (functions) that implement subscripting. The detailed API for subscript functions appears in src/include/nodes/subscripting.h. It may also be useful to read the array implementation in src/backend/utils/adt/arraysubs.c, or the simpler code in contrib/hstore/hstore_subs.c. Additional info appears in Array Types below. 👇

Okay so while the deets of the new type's internal representation are only known to the I/O functions and the other functions you create to work with the type, there are several properties of the internal representation that must be declared to PostgreSQL. Foremost of these is internallength. 🎀 Base data types can be:

  • fixed-length — where internallength is a positive integer, or
  • variable-length — indicated by setting internallength to VARIABLE. (Internally, this is represented by setting typlen to -1.)

The internal representation of all variable-length types must start with a 4-byte integer giving the total length of this value of the type. (Note that the length field is often encoded, as described in Section 73.2; it's unwise to access it directly. 🙈)

The optional flag PASSEDBYVALUE indicates that values of this data type are passed by value, rather than by reference. Types passed by value must be fixed-length, and their internal representation cannot be larger than the size of the Datum type (4 bytes on some machines, 8 bytes on others).

The alignment parameter specifies the storage alignment required for the data type. The allowed values equate to alignment on 1, 2, 4, or 8 byte boundaries. Note that variable-length types must have an alignment of at least 4, since they necessarily contain an int4 as their first component. ⭐

The storage parameter lets you pick storage strategies for variable-length data types. (Only plain is allowed for fixed-length types.) Here's the tea ☕:

  • plain — specifies that data of the type will always be stored in-line and not compressed.
  • extended — specifies that the system will first try to compress a long data value, and will move the value out of the main table row if it's still too long.
  • external — allows the value to be moved out of the main table, but the system will not try to compress it.
  • main — allows compression, but discourages moving the value out of the main table. (Data items with this storage strategy might still be moved out of the main table if there's no other way to make a row fit, but they'll be kept in the main table preferentially over extended and external items.)

All storage values other than plain imply that the functions of the data type can handle values that have been toasted, as described in Section 73.2 and Section 38.13.1. 🔥 The specific other value you give merely determines the default TOAST storage strategy for columns of a toastable data type; users can pick other strategies for individual columns using ALTER TABLE SET STORAGE.

The like_type parameter gives you an alternative method for specifying the basic representation properties of a data type: just copy them from some existing type, bestie. 🫶 The values of internallength, passedbyvalue, alignment, and storage are copied from the named type. (It is possible, though usually undesirable, to override some of these values by specifying them along with the LIKE clause.) Specifying representation this way is especially useful when the low-level implementation of the new type "piggybacks" on an existing type in some fashion.

The category and preferred parameters can be used to help control which implicit cast will be applied in ambiguous situations. Each data type belongs to a category named by a single ASCII character, and each type is either "preferred" or not within its category. The parser will prefer casting to preferred types (but only from other types within the same category) when this rule is helpful in resolving overloaded functions or operators. For more details see Chapter 10. 📚

For types that have no implicit casts to or from any other types, it's sufficient to leave these settings at the defaults. However, for a group of related types that have implicit casts, it's often helpful to mark them all as belonging to a category and select one or two of the "most general" types as being preferred within the category. The category parameter is especially useful when adding a user-defined type to an existing built-in category, such as the numeric or string types. But it's also possible to create new entirely-user-defined type categories — just select any ASCII character other than an upper-case letter to name such a category. 👑

A default value can be specified, in case a user wants columns of the data type to default to something other than the null value. Specify the default with the DEFAULT key word. (Such a default can be overridden by an explicit DEFAULT clause attached to a particular column.) 💖

To indicate that a type is a fixed-length array type, specify the type of the array elements using the ELEMENT key word. For example, to define an array of 4-byte integers (int4), specify ELEMENT = int4. For more details, see Array Types below.

To indicate the delimiter to be used between values in the external representation of arrays of this type, delimiter can be set to a specific character. The default delimiter is the comma (,). Note that the delimiter is associated with the array element type, not the array type itself. 🍓

If the optional Boolean parameter collatable is true, column definitions and expressions of the type may carry collation information through use of the COLLATE clause. It's up to the implementations of the functions operating on the type to actually make use of the collation information — this does not happen automatically merely by marking the type collatable. 😭

🌸 Array Types 🌸

Whenever a user-defined type is created, PostgreSQL automatically creates an associated array type, whose name consists of the element type's name prepended with an underscore, and truncated if necessary to keep it less than NAMEDATALEN bytes long. (If the name so generated collides with an existing type name, the process is repeated until a non-colliding name is found.) This implicitly-created array type is variable length and uses the built-in input and output functions array_in and array_out. Furthermore, this type is what the system uses for constructs such as ARRAY[] over the user-defined type. The array type tracks any changes in its element type's owner or schema, and is dropped if the element type is. 🦋

You might reasonably ask why there even is an ELEMENT option, if the system makes the correct array type automatically. The main case where it's useful to use ELEMENT is when you're making a fixed-length type that happens to be internally an array of a number of identical things, and you want to allow these things to be accessed directly by subscripting, in addition to whatever operations you plan to provide for the type as a whole. For example, type point is represented as just two floating-point numbers, which can be accessed using point[0] and point[1]. 👀

Note that this facility only works for fixed-length types whose internal form is exactly a sequence of identical fixed-length fields. For historical reasons (i.e., this is clearly wrong but it's far too late to change it 💀), subscripting of fixed-length array types starts from zero, rather than from one as for variable-length arrays.

Specifying the SUBSCRIPT option allows a data type to be subscripted, even though the system doesn't otherwise regard it as an array type. The behavior just described for fixed-length arrays is actually implemented by the SUBSCRIPT handler function raw_array_subscript_handler, which is used automatically if you specify ELEMENT for a fixed-length type without also writing SUBSCRIPT.

When specifying a custom SUBSCRIPT function, it's not necessary to specify ELEMENT unless the SUBSCRIPT handler function needs to consult typelem to find out what to return. Be aware that specifying ELEMENT causes the system to assume that the new type contains, or is somehow physically dependent on, the element type; thus for example changing properties of the element type won't be allowed if there are any columns of the dependent type. 🙅‍♀️

💖 PARAMETERS 💖

okay bestie, these are all the little pieces you can name when you're building a type. no skipping, i promise 🫶

  • name — the name (optionally schema-qualified!) of the type you're creating. yes, schema-qualified is allowed, go off 👑
  • attribute_name — the name of an attribute (aka a column) for the composite type.
  • data_type — the name of an existing data type that's gonna become a column of the composite type.
  • collation — the name of an existing collation to associate with a column of a composite type, or with a range type.
  • label — a string literal representing the textual label attached to one value of an enum type.
  • subtype — the name of the element type that the range type will represent ranges of.
  • subtype_operator_class — the name of a b-tree operator class for the subtype.
  • canonical_function — the name of the canonicalization function for the range type.
  • subtype_diff_function — the name of a difference function for the subtype.
  • multirange_type_name — the name of the corresponding multirange type.
  • input_function — the name of a function that converts data from the type's external textual form to its internal form.
  • output_function — the name of a function that converts data from the type's internal form to its external textual form.
  • receive_function — the name of a function that converts data from the type's external binary form to its internal form.
  • send_function — the name of a function that converts data from the type's internal form to its external binary form.
  • type_modifier_input_function — the name of a function that converts an array of modifier(s) for the type into internal form.
  • type_modifier_output_function — the name of a function that converts the internal form of the type's modifier(s) to external textual form.
  • analyze_function — the name of a function that performs statistical analysis for the data type.
  • subscript_function — the name of a function that defines what subscripting a value of the data type does.
  • internallength — a numeric constant that specifies the length in bytes of the new type's internal representation. the default assumption is that it's variable-length.
  • alignment — the storage alignment requirement of the data type. if specified, it must be char, int2, int4, or double; the default is int4.
  • storage — the storage strategy for the data type. if specified, it must be plain, external, extended, or main; the default is plain.
  • like_type — the name of an existing data type that the new type will have the same representation as. the values of internallength, passedbyvalue, alignment, and storage are copied from that type, unless they're overridden by explicit specification elsewhere in this CREATE TYPE command.
  • category — the category code (a single ASCII character) for this type. the default is 'U' for “user-defined type”. other standard category codes can be found in Table 53.65. you may also choose other ASCII characters in order to create custom categories 💅
  • preferred — true if this type is a preferred type within its type category, else false. the default is false. be very careful about creating a new preferred type within an existing type category, because this could cause surprising changes in behavior 😭
  • default — the default value for the data type. if this is omitted, the default is null.
  • element — the type being created is an array; this specifies the type of the array elements.
  • delimiter — the delimiter character to be used between values in arrays made of this type.
  • collatable — true if this type's operations can use collation information. the default is false.

🌸 NOTES 🌸

okay so here's the tea, and it's important tea 🍓

because there are no restrictions on use of a data type once it's been created, creating a base type or range type is tantamount to granting public execute permission on the functions mentioned in the type definition. usually this is not an issue for the sorts of functions that are useful in a type definition. but you might want to think twice before designing a type in a way that would require “secret” information to be used while converting it to or from external form 👀

before PostgreSQL version 8.3, the name of a generated array type was always exactly the element type's name with one underscore character (_) prepended. type names were therefore restricted in length to one fewer character than other names, lowkey tragic. while this is still usually the case, the array type name may vary from this in case of maximum-length names or collisions with user type names that begin with underscore. writing code that depends on this convention is therefore deprecated. instead, use pg_type.typarray to locate the array type associated with a given type. periodt ✨

it may be advisable to avoid using type and table names that begin with underscore. while the server will change generated array type names to avoid collisions with user-given names, there is still risk of confusion, particularly with old client software that may assume that type names beginning with underscores always represent arrays. not the vibe, honestly 🙈

before PostgreSQL version 8.2, the shell-type creation syntax CREATE TYPE name did not exist. the way to create a new base type was to create its input function first. in this approach, PostgreSQL will first see the name of the new data type as the return type of the input function. the shell type is implicitly created in this situation, and then it can be referenced in the definitions of the remaining I/O functions. this approach still works, but it's deprecated and might be disallowed in some future release. also, to avoid accidentally cluttering the catalogs with shell types as a result of simple typos in function definitions, a shell type will only be made this way when the input function is written in C. scary but valid 💀

in PostgreSQL version 16 and later, it's desirable for base types' input functions to return “soft” errors using the new errsave()/ereturn() mechanism, rather than throwing ereport() exceptions as in previous versions. see src/backend/utils/fmgr/README for more information. yes bestie, read the README, it's giving main character documentation 🌈

✨ EXAMPLES ✨

this example creates a composite type and uses it in a function definition. look at her go 💖

CREATE TYPE compfoo AS (f1 int, f2 text);

CREATE FUNCTION getfoo() RETURNS SETOF compfoo AS $$
    SELECT fooid, fooname FROM foo
$$ LANGUAGE SQL;

this example creates an enumerated type and uses it in a table definition. enum girlies rise 🎀

CREATE TYPE bug_status AS ENUM ('new', 'open', 'closed');

CREATE TABLE bug (
    id serial,
    description text,
    status bug_status
);

this example creates a range type:

CREATE TYPE float8_range AS RANGE (subtype = float8, subtype_diff = float8mi);

this example creates the base data type box and then uses the type in a table definition:

CREATE TYPE box;

CREATE FUNCTION my_box_in_function(cstring) RETURNS box AS ... ;
CREATE FUNCTION my_box_out_function(box) RETURNS cstring AS ... ;

CREATE TYPE box (
    INTERNALLENGTH = 16,
    INPUT = my_box_in_function,
    OUTPUT = my_box_out_function
);

CREATE TABLE myboxes (
    id integer,
    description box
);

if the internal structure of box were an array of four float4 elements, we might instead use:

CREATE TYPE box (
    INTERNALLENGTH = 16,
    INPUT = my_box_in_function,
    OUTPUT = my_box_out_function,
    ELEMENT = float4
);

which would allow a box value's component numbers to be accessed by subscripting. otherwise the type behaves the same as before. cute, no notes 🦋

this example creates a large object type and uses it in a table definition:

CREATE TYPE bigobj (
    INPUT = lo_filein, OUTPUT = lo_fileout,
    INTERNALLENGTH = VARIABLE
);
CREATE TABLE big_objs (
    id integer,
    obj bigobj
);

more examples, including suitable input and output functions, are in Section 38.13. go read, bestie, it's worth it 📚✨

💫 COMPATIBILITY 💫

the first form of the CREATE TYPE command, which creates a composite type, conforms to the SQL standard. the other forms are PostgreSQL extensions. the CREATE TYPE statement in the SQL standard also defines other forms that are not implemented in PostgreSQL. standard says one thing, PostgreSQL says “and also this” 💅

the ability to create a composite type with zero attributes is a PostgreSQL-specific deviation from the standard (analogous to the same case in CREATE TABLE). zero attributes? bold. iconic. very PostgreSQL 🌟

🦋 SEE ALSO 🦋

go forth and type responsibly, bestie. links in bio, etc. 🔥

this page is serving PostgreSQL 16.15, 2026, CREATE TYPE(7) realness — mwah 💋💖

📦 where this page comes from