Column INSERT/UPDATE Defaults

Column INSERT and UPDATE defaults refer to functions that create a default value for a particular column in a row as an INSERT or UPDATE statement is proceeding against that row, in the case where no value was provided to the INSERT or UPDATE statement for that column. That is, if a table has a column called “timestamp”, and an INSERT statement proceeds which does not include a value for this column, an INSERT default would create a new value, such as the current time, that is used as the value to be INSERTed into the “timestamp” column. If the statement does include a value for this column, then the default does not take place.

Column defaults can be server-side functions or constant values which are defined in the database along with the schema in DDL, or as SQL expressions which are rendered directly within an INSERT or UPDATE statement emitted by SQLAlchemy; they may also be client-side Python functions or constant values which are invoked by SQLAlchemy before data is passed to the database.


A column default handler should not be confused with a construct that intercepts and modifies incoming values for INSERT and UPDATE statements which are provided to the statement as it is invoked. This is known as data marshalling, where a column value is modified in some way by the application before being sent to the database. SQLAlchemy provides a few means of achieving this which include using custom datatypes, SQL execution events and in the ORM custom validators as well as attribute events. Column defaults are only invoked when there is no value present for a column in a SQL DML statement.

SQLAlchemy provides an array of features regarding default generation functions which take place for non-present values during INSERT and UPDATE statements. Options include:

  • Scalar values used as defaults during INSERT and UPDATE operations

  • Python functions which execute upon INSERT and UPDATE operations

  • SQL expressions which are embedded in INSERT statements (or in some cases execute beforehand)

  • SQL expressions which are embedded in UPDATE statements

  • Server side default values used during INSERT

  • Markers for server-side triggers used during UPDATE

The general rule for all insert/update defaults is that they only take effect if no value for a particular column is passed as an execute() parameter; otherwise, the given value is used.

Scalar Defaults

The simplest kind of default is a scalar value used as the default value of a column:

Table("mytable", meta,
    Column("somecolumn", Integer, default=12)

Above, the value “12” will be bound as the column value during an INSERT if no other value is supplied.

A scalar value may also be associated with an UPDATE statement, though this is not very common (as UPDATE statements are usually looking for dynamic defaults):

Table("mytable", meta,
    Column("somecolumn", Integer, onupdate=25)

Python-Executed Functions

The :paramref:`.Column.default` and :paramref:`.Column.onupdate` keyword arguments also accept Python functions. These functions are invoked at the time of insert or update if no other value for that column is supplied, and the value returned is used for the column’s value. Below illustrates a crude “sequence” that assigns an incrementing counter to a primary key column:

# a function which counts upwards
i = 0
def mydefault():
    global i
    i += 1
    return i

t = Table("mytable", meta,
    Column('id', Integer, primary_key=True, default=mydefault),

It should be noted that for real “incrementing sequence” behavior, the built-in capabilities of the database should normally be used, which may include sequence objects or other autoincrementing capabilities. For primary key columns, SQLAlchemy will in most cases use these capabilities automatically. See the API documentation for Column including the :paramref:`.Column.autoincrement` flag, as well as the section on Sequence later in this chapter for background on standard primary key generation techniques.

To illustrate onupdate, we assign the Python datetime function now to the :paramref:`.Column.onupdate` attribute:

import datetime

t = Table("mytable", meta,
    Column('id', Integer, primary_key=True),

    # define 'last_updated' to be populated with
    Column('last_updated', DateTime,,

When an update statement executes and no value is passed for last_updated, the Python function is executed and its return value used as the value for last_updated. Notice that we provide now as the function itself without calling it (i.e. there are no parenthesis following) - SQLAlchemy will execute the function at the time the statement executes.

Context-Sensitive Default Functions

The Python functions used by :paramref:`.Column.default` and :paramref:`.Column.onupdate` may also make use of the current statement’s context in order to determine a value. The context of a statement is an internal SQLAlchemy object which contains all information about the statement being executed, including its source expression, the parameters associated with it and the cursor. The typical use case for this context with regards to default generation is to have access to the other values being inserted or updated on the row. To access the context, provide a function that accepts a single context argument:

def mydefault(context):
    return context.get_current_parameters()['counter'] + 12

t = Table('mytable', meta,
    Column('counter', Integer),
    Column('counter_plus_twelve', Integer, default=mydefault, onupdate=mydefault)

The above default generation function is applied so that it will execute for all INSERT and UPDATE statements where a value for counter_plus_twelve was otherwise not provided, and the value will be that of whatever value is present in the execution for the counter column, plus the number 12.

For a single statement that is being executed using “executemany” style, e.g. with multiple parameter sets passed to Connection.execute(), the user- defined function is called once for each set of parameters. For the use case of a multi-valued Insert construct (e.g. with more than one VALUES clause set up via the Insert.values() method), the user-defined function is also called once for each set of parameters.

When the function is invoked, the special method DefaultExecutionContext.get_current_parameters() is available from the context object (an subclass of DefaultExecutionContext). This method returns a dictionary of column-key to values that represents the full set of values for the INSERT or UPDATE statement. In the case of a multi-valued INSERT construct, the subset of parameters that corresponds to the individual VALUES clause is isolated from the full parameter dictionary and returned alone.

New in version 1.2: Added DefaultExecutionContext.get_current_parameters() method, which improves upon the still-present DefaultExecutionContext.current_parameters attribute by offering the service of organizing multiple VALUES clauses into individual parameter dictionaries.

Client-Invoked SQL Expressions

The :paramref:`.Column.default` and :paramref:`.Column.onupdate` keywords may also be passed SQL expressions, which are in most cases rendered inline within the INSERT or UPDATE statement:

t = Table("mytable", meta,
    Column('id', Integer, primary_key=True),

    # define 'create_date' to default to now()
    Column('create_date', DateTime,,

    # define 'key' to pull its default from the 'keyvalues' table
    Column('key', String(20), default=select([keyvalues.c.key]).where(keyvalues.c.type='type1')),

    # define 'last_modified' to use the current_timestamp SQL function on update
    Column('last_modified', DateTime, onupdate=func.utc_timestamp())

Above, the create_date column will be populated with the result of the now() SQL function (which, depending on backend, compiles into NOW() or CURRENT_TIMESTAMP in most cases) during an INSERT statement, and the key column with the result of a SELECT subquery from another table. The last_modified column will be populated with the value of the SQL UTC_TIMESTAMP() MySQL function when an UPDATE statement is emitted for this table.


When using SQL functions with the func construct, we “call” the named function, e.g. with parenthesis as in This differs from when we specify a Python callable as a default such as datetime.datetime, where we pass the function itself, but we don’t invoke it ourselves. In the case of a SQL function, invoking returns the SQL expression object that will render the “NOW” function into the SQL being emitted.

Default and update SQL expressions specified by :paramref:`.Column.default` and :paramref:`.Column.onupdate` are invoked explicitly by SQLAlchemy when an INSERT or UPDATE statement occurs, typically rendered inline within the DML statement except in certain cases listed below. This is different than a “server side” default, which is part of the table’s DDL definition, e.g. as part of the “CREATE TABLE” statement, which are likely more common. For server side defaults, see the next section Server-invoked DDL-Explicit Default Expressions.

When a SQL expression indicated by :paramref:`.Column.default` is used with primary key columns, there are some cases where SQLAlchemy must “pre-execute” the default generation SQL function, meaning it is invoked in a separate SELECT statement, and the resulting value is passed as a parameter to the INSERT. This only occurs for primary key columns for an INSERT statement that is being asked to return this primary key value, where RETURNING or cursor.lastrowid may not be used. An Insert construct that specifies the :paramref:`~.expression.insert.inline` flag will always render default expressions inline.

When the statement is executed with a single set of parameters (that is, it is not an “executemany” style execution), the returned ResultProxy will contain a collection accessible via ResultProxy.postfetch_cols() which contains a list of all Column objects which had an inline-executed default. Similarly, all parameters which were bound to the statement, including all Python and SQL expressions which were pre-executed, are present in the ResultProxy.last_inserted_params() or ResultProxy.last_updated_params() collections on ResultProxy. The ResultProxy.inserted_primary_key collection contains a list of primary key values for the row inserted (a list so that single-column and composite- column primary keys are represented in the same format).

Server-invoked DDL-Explicit Default Expressions

A variant on the SQL expression default is the :paramref:`.Column.server_default`, which gets placed in the CREATE TABLE statement during a Table.create() operation:

t = Table('test', meta,
    Column('abc', String(20), server_default='abc'),
    Column('created_at', DateTime, server_default=func.sysdate()),
    Column('index_value', Integer, server_default=text("0"))

A create call for the above table will produce:

    abc varchar(20) default 'abc',
    created_at datetime default sysdate,
    index_value integer default 0

The above example illustrates the two typical use cases for :paramref:`.Column.server_default`, that of the SQL function (SYSDATE in the above example) as well as a server-side constant value (the integer “0” in the above example). It is advisable to use the text() construct for any literal SQL values as opposed to passing the raw value, as SQLAlchemy does not typically perform any quoting or escaping on these values.

Like client-generated expressions, :paramref:`.Column.server_default` can accommodate SQL expressions in general, however it is expected that these will usually be simple functions and expressions, and not the more complex cases like an embedded SELECT.

Marking Implicitly Generated Values, timestamps, and Triggered Columns

Columns which generate a new value on INSERT or UPDATE based on other server-side database mechanisms, such as database-specific auto-generating behaviors such as seen with TIMESTAMP columns on some platforms, as well as custom triggers that invoke upon INSERT or UPDATE to generate a new value, may be called out using FetchedValue as a marker:

t = Table('test', meta,
    Column('id', Integer, primary_key=True),
    Column('abc', TIMESTAMP, server_default=FetchedValue()),
    Column('def', String(20), server_onupdate=FetchedValue())

The FetchedValue indicator does not affect the rendered DDL for the CREATE TABLE. Instead, it marks the column as one that will have a new value populated by the database during the process of an INSERT or UPDATE statement, and for supporting databases may be used to indicate that the column should be part of a RETURNING or OUTPUT clause for the statement. Tools such as the SQLAlchemy ORM then make use of this marker in order to know how to get at the value of the column after such an operation. In particular, the ValuesBase.return_defaults() method can be used with an Insert or Update construct to indicate that these values should be returned.

For details on using FetchedValue with the ORM, see Fetching Server-Generated Defaults.

Defining Sequences

SQLAlchemy represents database sequences using the Sequence object, which is considered to be a special case of “column default”. It only has an effect on databases which have explicit support for sequences, which currently includes PostgreSQL, Oracle, MariaDB 10.3 or greater, and Firebird. The Sequence object is otherwise ignored.

The Sequence may be placed on any column as a “default” generator to be used during INSERT operations, and can also be configured to fire off during UPDATE operations if desired. It is most commonly used in conjunction with a single integer primary key column:

table = Table("cartitems", meta,
        Sequence('cart_id_seq', metadata=meta), primary_key=True),
    Column("description", String(40)),
    Column("createdate", DateTime())

Where above, the table “cartitems” is associated with a sequence named “cart_id_seq”. When INSERT statements take place for “cartitems”, and no value is passed for the “cart_id” column, the “cart_id_seq” sequence will be used to generate a value. Typically, the sequence function is embedded in the INSERT statement, which is combined with RETURNING so that the newly generated value can be returned to the Python code:

INSERT INTO cartitems (cart_id, description, createdate)
VALUES (next_val(cart_id_seq), 'some description', '2015-10-15 12:00:15')

When the Sequence is associated with a Column as its Python-side default generator, the Sequence will also be subject to “CREATE SEQUENCE” and “DROP SEQUENCE” DDL when similar DDL is emitted for the owning Table. This is a limited scope convenience feature that does not accommodate for inheritance of other aspects of the MetaData, such as the default schema. Therefore, it is best practice that for a Sequence which is local to a certain Column / Table, that it be explicitly associated with the MetaData using the :paramref:`.Sequence.metadata` parameter. See the section Associating a Sequence with the MetaData for more background on this.

Associating a Sequence on a SERIAL column

PostgreSQL’s SERIAL datatype is an auto-incrementing type that implies the implicit creation of a PostgreSQL sequence when CREATE TABLE is emitted. If a Column specifies an explicit Sequence object which also specifies a true value for the :paramref:`.Sequence.optional` boolean flag, the Sequence will not take effect under PostgreSQL, and the SERIAL datatype will proceed normally. Instead, the Sequence will only take effect when used against other sequence-supporting databases, currently Oracle and Firebird.

Executing a Sequence Standalone

A SEQUENCE is a first class schema object in SQL and can be used to generate values independently in the database. If you have a Sequence object, it can be invoked with its “next value” instruction by passing it directly to a SQL execution method:

with my_engine.connect() as conn:
    seq = Sequence('some_sequence')
    nextid = conn.execute(seq)

In order to embed the “next value” function of a Sequence inside of a SQL statement like a SELECT or INSERT, use the Sequence.next_value() method, which will render at statement compilation time a SQL function that is appropriate for the target backend:

>>> my_seq = Sequence('some_sequence')
>>> stmt = select([my_seq.next_value()])
>>> print stmt.compile(dialect=postgresql.dialect())
SELECT nextval('some_sequence') AS next_value_1

Associating a Sequence with the MetaData

For many years, the SQLAlchemy documentation referred to the example of associating a Sequence with a table as follows:

table = Table("cartitems", meta,
    Column("cart_id", Integer, Sequence('cart_id_seq'),
    Column("description", String(40)),
    Column("createdate", DateTime())

While the above is a prominent idiomatic pattern, it is recommended that the Sequence in most cases be explicitly associated with the MetaData, using the :paramref:`.Sequence.metadata` parameter:

table = Table("cartitems", meta,
        Sequence('cart_id_seq', metadata=meta), primary_key=True),
    Column("description", String(40)),
    Column("createdate", DateTime())

The Sequence object is a first class schema construct that can exist independently of any table in a database, and can also be shared among tables. Therefore SQLAlchemy does not implicitly modify the Sequence when it is associated with a Column object as either the Python-side or server-side default generator. While the CREATE SEQUENCE / DROP SEQUENCE DDL is emitted for a Sequence defined as a Python side generator at the same time the table itself is subject to CREATE or DROP, this is a convenience feature that does not imply that the Sequence is fully associated with the MetaData object.

Explicitly associating the Sequence with MetaData allows for the following behaviors:

Since the vast majority of cases that deal with Sequence expect that Sequence to be fully “owned” by the associated Table and that options like default schema are propagated, setting the :paramref:`.Sequence.metadata` parameter should be considered a best practice.

Associating a Sequence as the Server Side Default


The following technique is known to work only with the PostgreSQL database. It does not work with Oracle.

The preceding sections illustrate how to associate a Sequence with a Column as the Python side default generator:

    "cart_id", Integer, Sequence('cart_id_seq', metadata=meta),

In the above case, the Sequence will automatically be subject to CREATE SEQUENCE / DROP SEQUENCE DDL when the related Table is subject to CREATE / DROP. However, the sequence will not be present as the server-side default for the column when CREATE TABLE is emitted.

If we want the sequence to be used as a server-side default, meaning it takes place even if we emit INSERT commands to the table from the SQL command line, we can use the :paramref:`.Column.server_default` parameter in conjunction with the value-generation function of the sequence, available from the Sequence.next_value() method. Below we illustrate the same Sequence being associated with the Column both as the Python-side default generator as well as the server-side default generator:

cart_id_seq = Sequence('cart_id_seq', metadata=meta)
table = Table("cartitems", meta,
        "cart_id", Integer, cart_id_seq,
        server_default=cart_id_seq.next_value(), primary_key=True),
    Column("description", String(40)),
    Column("createdate", DateTime())

or with the ORM:

class CartItem(Base):
    __tablename__ = 'cartitems'

    cart_id_seq = Sequence('cart_id_seq', metadata=Base.metadata)
    cart_id = Column(
        Integer, cart_id_seq,
        server_default=cart_id_seq.next_value(), primary_key=True)
    description = Column(String(40))
    createdate = Column(DateTime)

When the “CREATE TABLE” statement is emitted, on PostgreSQL it would be emitted as:

CREATE TABLE cartitems (
    cart_id INTEGER DEFAULT nextval('cart_id_seq') NOT NULL,
    description VARCHAR(40),
    PRIMARY KEY (cart_id)

Placement of the Sequence in both the Python-side and server-side default generation contexts ensures that the “primary key fetch” logic works in all cases. Typically, sequence-enabled databases also support RETURNING for INSERT statements, which is used automatically by SQLAlchemy when emitting this statement. However if RETURNING is not used for a particular insert, then SQLAlchemy would prefer to “pre-execute” the sequence outside of the INSERT statement itself, which only works if the sequence is included as the Python-side default generator function.

The example also associates the Sequence with the enclosing MetaData directly, which again ensures that the Sequence is fully associated with the parameters of the MetaData collection including the default schema, if any.

See also

Sequences/SERIAL/IDENTITY - in the PostgreSQL dialect documentation

RETURNING Support - in the Oracle dialect documentation

Computed (GENERATED ALWAYS AS) Columns

New in version 1.3.11.

The Computed construct allows a Column to be declared in DDL as a “GENERATED ALWAYS AS” column, that is, one which has a value that is computed by the database server. The construct accepts a SQL expression typically declared textually using a string or the text() construct, in a similar manner as that of CheckConstraint. The SQL expression is then interpreted by the database server in order to determine the value for the column within a row.


from sqlalchemy import Table, Column, MetaData, Integer, Computed

metadata = MetaData()

square = Table(
    Column("id", Integer, primary_key=True),
    Column("side", Integer),
    Column("area", Integer, Computed("side * side")),
    Column("perimeter", Integer, Computed("4 * side")),

The DDL for the square table when run on a PostgreSQL 12 backend will look like:

    side INTEGER,
    perimeter INTEGER GENERATED ALWAYS AS (4 * side) STORED,
    PRIMARY KEY (id)

Whether the value is persisted upon INSERT and UPDATE, or if it is calculated on fetch, is an implementation detail of the database; the former is known as “stored” and the latter is known as “virtual”. Some database implementations support both, but some only support one or the other. The optional :paramref:`.Computed.persisted` flag may be specified as True or False to indicate if the “STORED” or “VIRTUAL” keyword should be rendered in DDL, however this will raise an error if the keyword is not supported by the target backend; leaving it unset will use a working default for the target backend.

The Computed construct is a subclass of the FetchedValue object, and will set itself up as both the “server default” and “server onupdate” generator for the target Column, meaning it will be treated as a default generating column when INSERT and UPDATE statements are generated, as well as that it will be fetched as a generating column when using the ORM. This includes that it will be part of the RETURNING clause of the database for databases which support RETURNING and the generated values are to be eagerly fetched.


A Column that is defined with the Computed construct may not store any value outside of that which the server applies to it; SQLAlchemy’s behavior when a value is passed for such a column to be written in INSERT or UPDATE is currently that the value will be ignored.

“GENERATED ALWAYS AS” is currently known to be supported by:

  • MySQL version 5.7 and onwards

  • MariaDB 10.x series and onwards

  • PostgreSQL as of version 12

  • Oracle - with the caveat that RETURNING does not work correctly with UPDATE (a warning will be emitted to this effect when the UPDATE..RETURNING that includes a computed column is rendered)

  • Microsoft SQL Server

  • Firebird

When Computed is used with an unsupported backend, if the target dialect does not support it, a CompileError is raised when attempting to render the construct. Otherwise, if the dialect supports it but the particular database server version in use does not, then a subclass of DBAPIError, usually OperationalError, is raised when the DDL is emitted to the database.

See also


Default Objects API

class sqlalchemy.schema.Computed(sqltext, persisted=None)

Defines a generated column, i.e. “GENERATED ALWAYS AS” syntax.

The Computed construct is an inline construct added to the argument list of a Column object:

from sqlalchemy import Computed

Table('square', meta,
    Column('side', Float, nullable=False),
    Column('area', Float, Computed('side * side'))

See the linked documentation below for complete details.

New in version 1.3.11.

class sqlalchemy.schema.ColumnDefault(arg, **kwargs)

A plain default value on a column.

This could correspond to a constant, a callable function, or a SQL clause.

ColumnDefault is generated automatically whenever the default, onupdate arguments of Column are used. A ColumnDefault can be passed positionally as well.

For example, the following:

Column('foo', Integer, default=50)

Is equivalent to:

Column('foo', Integer, ColumnDefault(50))
class sqlalchemy.schema.DefaultClause(arg, for_update=False, _reflected=False)

A DDL-specified DEFAULT column value.

DefaultClause is a FetchedValue that also generates a “DEFAULT” clause when “CREATE TABLE” is emitted.

DefaultClause is generated automatically whenever the server_default, server_onupdate arguments of Column are used. A DefaultClause can be passed positionally as well.

For example, the following:

Column('foo', Integer, server_default="50")

Is equivalent to:

Column('foo', Integer, DefaultClause("50"))
class sqlalchemy.schema.DefaultGenerator(for_update=False)

Base class for column default values.

class sqlalchemy.schema.FetchedValue(for_update=False)

A marker for a transparent database-side default.

Use FetchedValue when the database is configured to provide some automatic default for a column.


Column('foo', Integer, FetchedValue())

Would indicate that some trigger or default generator will create a new value for the foo column during an INSERT.

class sqlalchemy.schema.Sequence(name, start=None, increment=None, minvalue=None, maxvalue=None, nominvalue=None, nomaxvalue=None, cycle=None, schema=None, cache=None, order=None, optional=False, quote=None, metadata=None, quote_schema=None, for_update=False)

Represents a named database sequence.

The Sequence object represents the name and configurational parameters of a database sequence. It also represents a construct that can be “executed” by a SQLAlchemy Engine or Connection, rendering the appropriate “next value” function for the target database and returning a result.

The Sequence is typically associated with a primary key column:

some_table = Table(
    'some_table', metadata,
    Column('id', Integer, Sequence('some_table_seq'),

When CREATE TABLE is emitted for the above Table, if the target platform supports sequences, a CREATE SEQUENCE statement will be emitted as well. For platforms that don’t support sequences, the Sequence construct is ignored.

property bind

Return the connectable associated with this default.

create(bind=None, checkfirst=True)

Creates this sequence in the database.

drop(bind=None, checkfirst=True)

Drops this sequence from the database.


Return a next_value function element which will render the appropriate increment function for this Sequence within any SQL expression.