Flask-SQLAlchemy¶

Flask-SQLAlchemy is an extension for Flask that adds support for SQLAlchemy to your application. It aims to simplify using SQLAlchemy with Flask by providing useful defaults and extra helpers that make it easier to accomplish common tasks.

See the SQLAlchemy documentation to learn how to work with the ORM in depth. The following documentation is a brief overview of the most common tasks, as well as the features specific to Flask-SQLAlchemy.

Requirements¶

Our Version	Python	Flask	SQLAlchemy
2.x	2.7, 3.4+	0.12+	0.8+ or 1.0.10+ w/ Python 3.7
3.0+ (in dev)	2.7, 3.5+	1.0+	1.0+

User Guide¶

Quickstart¶

Flask-SQLAlchemy is fun to use, incredibly easy for basic applications, and readily extends for larger applications. For the complete guide, checkout the API documentation on the SQLAlchemy class.

Installation¶

Install and update using pip:

$ pip install -U Flask-SQLAlchemy

A Minimal Application¶

For the common case of having one Flask application all you have to do is to create your Flask application, load the configuration of choice and then create the SQLAlchemy object by passing it the application.

Once created, that object then contains all the functions and helpers from both sqlalchemy and sqlalchemy.orm. Furthermore it provides a class called Model that is a declarative base which can be used to declare models:

from flask import Flask
from flask_sqlalchemy import SQLAlchemy

app = Flask(__name__)
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:////tmp/test.db'
db = SQLAlchemy(app)


class User(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    username = db.Column(db.String(80), unique=True, nullable=False)
    email = db.Column(db.String(120), unique=True, nullable=False)

    def __repr__(self):
        return '<User %r>' % self.username

To create the initial database, just import the db object from an interactive Python shell and run the SQLAlchemy.create_all() method to create the tables and database:

>>> from yourapplication import db
>>> db.create_all()

Boom, and there is your database. Now to create some users:

>>> from yourapplication import User
>>> admin = User(username='admin', email='admin@example.com')
>>> guest = User(username='guest', email='guest@example.com')

But they are not yet in the database, so let’s make sure they are:

>>> db.session.add(admin)
>>> db.session.add(guest)
>>> db.session.commit()

Accessing the data in database is easy as a pie:

>>> User.query.all()
[<User u'admin'>, <User u'guest'>]
>>> User.query.filter_by(username='admin').first()
<User u'admin'>

Note how we never defined a __init__ method on the User class? That’s because SQLAlchemy adds an implicit constructor to all model classes which accepts keyword arguments for all its columns and relationships. If you decide to override the constructor for any reason, make sure to keep accepting **kwargs and call the super constructor with those **kwargs to preserve this behavior:

class Foo(db.Model):
    # ...
    def __init__(self, **kwargs):
        super(Foo, self).__init__(**kwargs)
        # do custom stuff

Simple Relationships¶

SQLAlchemy connects to relational databases and what relational databases are really good at are relations. As such, we shall have an example of an application that uses two tables that have a relationship to each other:

from datetime import datetime


class Post(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    title = db.Column(db.String(80), nullable=False)
    body = db.Column(db.Text, nullable=False)
    pub_date = db.Column(db.DateTime, nullable=False,
        default=datetime.utcnow)

    category_id = db.Column(db.Integer, db.ForeignKey('category.id'),
        nullable=False)
    category = db.relationship('Category',
        backref=db.backref('posts', lazy=True))

    def __repr__(self):
        return '<Post %r>' % self.title


class Category(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    name = db.Column(db.String(50), nullable=False)

    def __repr__(self):
        return '<Category %r>' % self.name

First let’s create some objects:

>>> py = Category(name='Python')
>>> Post(title='Hello Python!', body='Python is pretty cool', category=py)
>>> p = Post(title='Snakes', body='Ssssssss')
>>> py.posts.append(p)
>>> db.session.add(py)

As you can see, there is no need to add the Post objects to the session. Since the Category is part of the session all objects associated with it through relationships will be added too. It does not matter whether db.session.add() is called before or after creating these objects. The association can also be done on either side of the relationship - so a post can be created with a category or it can be added to the list of posts of the category.

Let’s look at the posts. Accessing them will load them from the database since the relationship is lazy-loaded, but you will probably not notice the difference - loading a list is quite fast:

>>> py.posts
[<Post 'Hello Python!'>, <Post 'Snakes'>]

While lazy-loading a relationship is fast, it can easily become a major bottleneck when you end up triggering extra queries in a loop for more than a few objects. For this case, SQLAlchemy lets you override the loading strategy on the query level. If you wanted a single query to load all categories and their posts, you could do it like this:

>>> from sqlalchemy.orm import joinedload
>>> query = Category.query.options(joinedload('posts'))
>>> for category in query:
...     print category, category.posts
<Category u'Python'> [<Post u'Hello Python!'>, <Post u'Snakes'>]

If you want to get a query object for that relationship, you can do so using with_parent(). Let’s exclude that post about Snakes for example:

>>> Post.query.with_parent(py).filter(Post.title != 'Snakes').all()
[<Post 'Hello Python!'>]

Road to Enlightenment¶

The only things you need to know compared to plain SQLAlchemy are:

SQLAlchemy gives you access to the following things:
- all the functions and classes from sqlalchemy and sqlalchemy.orm
- a preconfigured scoped session called session
- the metadata
- the engine
- a SQLAlchemy.create_all() and SQLAlchemy.drop_all() methods to create and drop tables according to the models.
- a Model baseclass that is a configured declarative base.
The Model declarative base class behaves like a regular Python class but has a query attribute attached that can be used to query the model. (Model and BaseQuery)
You have to commit the session, but you don’t have to remove it at the end of the request, Flask-SQLAlchemy does that for you.

Introduction into Contexts¶

If you are planning on using only one application you can largely skip this chapter. Just pass your application to the SQLAlchemy constructor and you’re usually set. However if you want to use more than one application or create the application dynamically in a function you want to read on.

If you define your application in a function, but the SQLAlchemy object globally, how does the latter learn about the former? The answer is the init_app() function:

from flask import Flask
from flask_sqlalchemy import SQLAlchemy

db = SQLAlchemy()

def create_app():
    app = Flask(__name__)
    db.init_app(app)
    return app

What it does is prepare the application to work with SQLAlchemy. However that does not now bind the SQLAlchemy object to your application. Why doesn’t it do that? Because there might be more than one application created.

So how does SQLAlchemy come to know about your application? You will have to setup an application context. If you are working inside a Flask view function or a CLI command, that automatically happens. However, if you are working inside the interactive shell, you will have to do that yourself (see Creating an Application Context).

If you try to perform database operations outside an application context, you will see the following error:

No application found. Either work inside a view function or push an application context.

In a nutshell, do something like this:

>>> from yourapp import create_app
>>> app = create_app()
>>> app.app_context().push()

Alternatively, use the with-statement to take care of setup and teardown:

def my_function():
    with app.app_context():
        user = db.User(...)
        db.session.add(user)
        db.session.commit()

Some functions inside Flask-SQLAlchemy also accept optionally the application to operate on:

>>> from yourapp import db, create_app
>>> db.create_all(app=create_app())

Configuration¶

The following configuration values exist for Flask-SQLAlchemy. Flask-SQLAlchemy loads these values from your main Flask config which can be populated in various ways. Note that some of those cannot be modified after the engine was created so make sure to configure as early as possible and to not modify them at runtime.

Configuration Keys¶

A list of configuration keys currently understood by the extension:

`SQLALCHEMY_DATABASE_URI`	The database URI that should be used for the connection. Examples: `sqlite:////tmp/test.db` `mysql://username:password@server/db`
`SQLALCHEMY_BINDS`	A dictionary that maps bind keys to SQLAlchemy connection URIs. For more information about binds see Multiple Databases with Binds.
`SQLALCHEMY_ECHO`	If set to True SQLAlchemy will log all the statements issued to stderr which can be useful for debugging.
`SQLALCHEMY_RECORD_QUERIES`	Can be used to explicitly disable or enable query recording. Query recording automatically happens in debug or testing mode. See `get_debug_queries()` for more information.
`SQLALCHEMY_NATIVE_UNICODE`	Can be used to explicitly disable native unicode support. This is required for some database adapters (like PostgreSQL on some Ubuntu versions) when used with improper database defaults that specify encoding-less databases. Deprecated as of v2.4 and will be removed in v3.0.
`SQLALCHEMY_POOL_SIZE`	The size of the database pool. Defaults to the engine’s default (usually 5). Deprecated as of v2.4 and will be removed in v3.0.
`SQLALCHEMY_POOL_TIMEOUT`	Specifies the connection timeout in seconds for the pool. Deprecated as of v2.4 and will be removed in v3.0.
`SQLALCHEMY_POOL_RECYCLE`	Number of seconds after which a connection is automatically recycled. This is required for MySQL, which removes connections after 8 hours idle by default. Note that Flask-SQLAlchemy automatically sets this to 2 hours if MySQL is used. Some backends may use a different default timeout value. For more information about timeouts see Timeouts. Deprecated as of v2.4 and will be removed in v3.0.
`SQLALCHEMY_MAX_OVERFLOW`	Controls the number of connections that can be created after the pool reached its maximum size. When those additional connections are returned to the pool, they are disconnected and discarded. Deprecated as of v2.4 and will be removed in v3.0.
`SQLALCHEMY_TRACK_MODIFICATIONS`	If set to `True`, Flask-SQLAlchemy will track modifications of objects and emit signals. The default is `None`, which enables tracking but issues a warning that it will be disabled by default in the future. This requires extra memory and should be disabled if not needed.
`SQLALCHEMY_ENGINE_OPTIONS`	A dictionary of keyword args to send to `create_engine()`. See also `engine_options` to `SQLAlchemy`.

New in version 0.8: The SQLALCHEMY_NATIVE_UNICODE, SQLALCHEMY_POOL_SIZE, SQLALCHEMY_POOL_TIMEOUT and SQLALCHEMY_POOL_RECYCLE configuration keys were added.

New in version 0.12: The SQLALCHEMY_BINDS configuration key was added.

New in version 0.17: The SQLALCHEMY_MAX_OVERFLOW configuration key was added.

New in version 2.0: The SQLALCHEMY_TRACK_MODIFICATIONS configuration key was added.

Changed in version 2.1: SQLALCHEMY_TRACK_MODIFICATIONS will warn if unset.

Changed in version 2.4: * SQLALCHEMY_ENGINE_OPTIONS configuration key was added. * Deprecated keys

SQLALCHEMY_NATIVE_UNICODE

SQLALCHEMY_POOL_SIZE

SQLALCHEMY_POOL_TIMEOUT

SQLALCHEMY_POOL_RECYCLE

SQLALCHEMY_MAX_OVERFLOW

Changed in version 2.4.3: Deprecated SQLALCHEMY_COMMIT_ON_TEARDOWN.

Connection URI Format¶

For a complete list of connection URIs head over to the SQLAlchemy documentation under (Supported Databases). This here shows some common connection strings.

SQLAlchemy indicates the source of an Engine as a URI combined with optional keyword arguments to specify options for the Engine. The form of the URI is:

dialect+driver://username:password@host:port/database

Many of the parts in the string are optional. If no driver is specified the default one is selected (make sure to not include the + in that case).

Postgres:

postgresql://scott:tiger@localhost/mydatabase

MySQL:

mysql://scott:tiger@localhost/mydatabase

Oracle:

oracle://scott:tiger@127.0.0.1:1521/sidname

SQLite (note that platform path conventions apply):

#Unix/Mac (note the four leading slashes)
sqlite:////absolute/path/to/foo.db
#Windows (note 3 leading forward slashes and backslash escapes)
sqlite:///C:\\absolute\\path\\to\\foo.db
#Windows (alternative using raw string)
r'sqlite:///C:\absolute\path\to\foo.db'

Using custom MetaData and naming conventions¶

You can optionally construct the SQLAlchemy object with a custom MetaData object. This allows you to, among other things, specify a custom constraint naming convention in conjunction with SQLAlchemy 0.9.2 or higher. Doing so is important for dealing with database migrations (for instance using alembic as stated here. Here’s an example, as suggested by the SQLAlchemy docs:

from sqlalchemy import MetaData
from flask import Flask
from flask_sqlalchemy import SQLAlchemy

convention = {
    "ix": 'ix_%(column_0_label)s',
    "uq": "uq_%(table_name)s_%(column_0_name)s",
    "ck": "ck_%(table_name)s_%(constraint_name)s",
    "fk": "fk_%(table_name)s_%(column_0_name)s_%(referred_table_name)s",
    "pk": "pk_%(table_name)s"
}

metadata = MetaData(naming_convention=convention)
db = SQLAlchemy(app, metadata=metadata)

For more info about MetaData, check out the official docs on it.

Timeouts¶

Certain database backends may impose different inactive connection timeouts, which interferes with Flask-SQLAlchemy’s connection pooling.

By default, MariaDB is configured to have a 600 second timeout. This often surfaces hard to debug, production environment only exceptions like 2013: Lost connection to MySQL server during query.

If you are using a backend (or a pre-configured database-as-a-service) with a lower connection timeout, it is recommended that you set SQLALCHEMY_POOL_RECYCLE to a value less than your backend’s timeout.

Declaring Models¶

Generally Flask-SQLAlchemy behaves like a properly configured declarative base from the declarative extension. As such we recommend reading the SQLAlchemy docs for a full reference. However the most common use cases are also documented here.

Things to keep in mind:

The baseclass for all your models is called db.Model. It’s stored on the SQLAlchemy instance you have to create. See Quickstart for more details.
Some parts that are required in SQLAlchemy are optional in Flask-SQLAlchemy. For instance the table name is automatically set for you unless overridden. It’s derived from the class name converted to lowercase and with “CamelCase” converted to “camel_case”. To override the table name, set the __tablename__ class attribute.

Simple Example¶

A very simple example:

class User(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    username = db.Column(db.String(80), unique=True, nullable=False)
    email = db.Column(db.String(120), unique=True, nullable=False)

    def __repr__(self):
        return '<User %r>' % self.username

Use Column to define a column. The name of the column is the name you assign it to. If you want to use a different name in the table you can provide an optional first argument which is a string with the desired column name. Primary keys are marked with primary_key=True. Multiple keys can be marked as primary keys in which case they become a compound primary key.

The types of the column are the first argument to Column. You can either provide them directly or call them to further specify them (like providing a length). The following types are the most common:

`Integer`	an integer
`String(size)`	a string with a maximum length (optional in some databases, e.g. PostgreSQL)
`Text`	some longer unicode text
`DateTime`	date and time expressed as Python `datetime` object.
`Float`	stores floating point values
`Boolean`	stores a boolean value
`PickleType`	stores a pickled Python object
`LargeBinary`	stores large arbitrary binary data

One-to-Many Relationships¶

The most common relationships are one-to-many relationships. Because relationships are declared before they are established you can use strings to refer to classes that are not created yet (for instance if Person defines a relationship to Address which is declared later in the file).

Relationships are expressed with the relationship() function. However the foreign key has to be separately declared with the ForeignKey class:

class Person(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    name = db.Column(db.String(50), nullable=False)
    addresses = db.relationship('Address', backref='person', lazy=True)

class Address(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    email = db.Column(db.String(120), nullable=False)
    person_id = db.Column(db.Integer, db.ForeignKey('person.id'),
        nullable=False)

What does db.relationship() do? That function returns a new property that can do multiple things. In this case we told it to point to the Address class and load multiple of those. How does it know that this will return more than one address? Because SQLAlchemy guesses a useful default from your declaration. If you would want to have a one-to-one relationship you can pass uselist=False to relationship().

Since a person with no name or an email address with no address associated makes no sense, nullable=False tells SQLAlchemy to create the column as NOT NULL. This is implied for primary key columns, but it’s a good idea to specify it for all other columns to make it clear to other people working on your code that you did actually want a nullable column and did not just forget to add it.

So what do backref and lazy mean? backref is a simple way to also declare a new property on the Address class. You can then also use my_address.person to get to the person at that address. lazy defines when SQLAlchemy will load the data from the database:

'select' / True (which is the default, but explicit is better than implicit) means that SQLAlchemy will load the data as necessary in one go using a standard select statement.
'joined' / False tells SQLAlchemy to load the relationship in the same query as the parent using a JOIN statement.
'subquery' works like 'joined' but instead SQLAlchemy will use a subquery.
'dynamic' is special and can be useful if you have many items and always want to apply additional SQL filters to them. Instead of loading the items SQLAlchemy will return another query object which you can further refine before loading the items. Note that this cannot be turned into a different loading strategy when querying so it’s often a good idea to avoid using this in favor of lazy=True. A query object equivalent to a dynamic user.addresses relationship can be created using Address.query.with_parent(user) while still being able to use lazy or eager loading on the relationship itself as necessary.

How do you define the lazy status for backrefs? By using the backref() function:

class Person(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    name = db.Column(db.String(50), nullable=False)
    addresses = db.relationship('Address', lazy='select',
        backref=db.backref('person', lazy='joined'))

Many-to-Many Relationships¶

If you want to use many-to-many relationships you will need to define a helper table that is used for the relationship. For this helper table it is strongly recommended to not use a model but an actual table:

tags = db.Table('tags',
    db.Column('tag_id', db.Integer, db.ForeignKey('tag.id'), primary_key=True),
    db.Column('page_id', db.Integer, db.ForeignKey('page.id'), primary_key=True)
)

class Page(db.Model):
    id = db.Column(db.Integer, primary_key=True)
    tags = db.relationship('Tag', secondary=tags, lazy='subquery',
        backref=db.backref('pages', lazy=True))

class Tag(db.Model):
    id = db.Column(db.Integer, primary_key=True)

Here we configured Page.tags to be loaded immediately after loading a Page, but using a separate query. This always results in two queries when retrieving a Page, but when querying for multiple pages you will not get additional queries.

The list of pages for a tag on the other hand is something that’s rarely needed. For example, you won’t need that list when retrieving the tags for a specific page. Therefore, the backref is set to be lazy-loaded so that accessing it for the first time will trigger a query to get the list of pages for that tag. If you need to apply further query options on that list, you could either switch to the 'dynamic' strategy - with the drawbacks mentioned above - or get a query object using Page.query.with_parent(some_tag) and then use it exactly as you would with the query object from a dynamic relationship.

Select, Insert, Delete¶

Now that you have declared models it’s time to query the data from the database. We will be using the model definitions from the Quickstart chapter.

Inserting Records¶

Before we can query something we will have to insert some data. All your models should have a constructor, so make sure to add one if you forgot. Constructors are only used by you, not by SQLAlchemy internally so it’s entirely up to you how you define them.

Inserting data into the database is a three step process:

Create the Python object
Add it to the session
Commit the session

The session here is not the Flask session, but the Flask-SQLAlchemy one. It is essentially a beefed up version of a database transaction. This is how it works:

>>> from yourapp import User
>>> me = User('admin', 'admin@example.com')
>>> db.session.add(me)
>>> db.session.commit()

Alright, that was not hard. What happens at what point? Before you add the object to the session, SQLAlchemy basically does not plan on adding it to the transaction. That is good because you can still discard the changes. For example think about creating the post at a page but you only want to pass the post to the template for preview rendering instead of storing it in the database.

The add() function call then adds the object. It will issue an INSERT statement for the database but because the transaction is still not committed you won’t get an ID back immediately. If you do the commit, your user will have an ID:

>>> me.id
1

Deleting Records¶

Deleting records is very similar, instead of add() use delete():

>>> db.session.delete(me)
>>> db.session.commit()

Querying Records¶

So how do we get data back out of our database? For this purpose Flask-SQLAlchemy provides a query attribute on your Model class. When you access it you will get back a new query object over all records. You can then use methods like filter() to filter the records before you fire the select with all() or first(). If you want to go by primary key you can also use get().

The following queries assume following entries in the database:

id	username	email
1	admin	admin@example.com
2	peter	peter@example.org
3	guest	guest@example.com

Retrieve a user by username:

>>> peter = User.query.filter_by(username='peter').first()
>>> peter.id
2
>>> peter.email
u'peter@example.org'

Same as above but for a non existing username gives None:

>>> missing = User.query.filter_by(username='missing').first()
>>> missing is None
True

Selecting a bunch of users by a more complex expression:

>>> User.query.filter(User.email.endswith('@example.com')).all()
[<User u'admin'>, <User u'guest'>]

Ordering users by something:

>>> User.query.order_by(User.username).all()
[<User u'admin'>, <User u'guest'>, <User u'peter'>]

Limiting users:

>>> User.query.limit(1).all()
[<User u'admin'>]

Getting user by primary key:

>>> User.query.get(1)
<User u'admin'>

Queries in Views¶

If you write a Flask view function it’s often very handy to return a 404 error for missing entries. Because this is a very common idiom, Flask-SQLAlchemy provides a helper for this exact purpose. Instead of get() one can use get_or_404() and instead of first() first_or_404(). This will raise 404 errors instead of returning None:

@app.route('/user/<username>')
def show_user(username):
    user = User.query.filter_by(username=username).first_or_404()
    return render_template('show_user.html', user=user)

Also, if you want to add a description with abort(), you can use it as argument as well.

>>> User.query.filter_by(username=username).first_or_404(description='There is no data with {}'.format(username))

Multiple Databases with Binds¶

Starting with 0.12 Flask-SQLAlchemy can easily connect to multiple databases. To achieve that it preconfigures SQLAlchemy to support multiple “binds”.

What are binds? In SQLAlchemy speak a bind is something that can execute SQL statements and is usually a connection or engine. In Flask-SQLAlchemy binds are always engines that are created for you automatically behind the scenes. Each of these engines is then associated with a short key (the bind key). This key is then used at model declaration time to assocate a model with a specific engine.

If no bind key is specified for a model the default connection is used instead (as configured by SQLALCHEMY_DATABASE_URI).

Example Configuration¶

The following configuration declares three database connections. The special default one as well as two others named users (for the users) and appmeta (which connects to a sqlite database for read only access to some data the application provides internally):

SQLALCHEMY_DATABASE_URI = 'postgres://localhost/main'
SQLALCHEMY_BINDS = {
    'users':        'mysqldb://localhost/users',
    'appmeta':      'sqlite:////path/to/appmeta.db'
}

Creating and Dropping Tables¶

The create_all() and drop_all() methods by default operate on all declared binds, including the default one. This behavior can be customized by providing the bind parameter. It takes either a single bind name, '__all__' to refer to all binds or a list of binds. The default bind (SQLALCHEMY_DATABASE_URI) is named None:

>>> db.create_all()
>>> db.create_all(bind=['users'])
>>> db.create_all(bind='appmeta')
>>> db.drop_all(bind=None)

Referring to Binds¶

If you declare a model you can specify the bind to use with the __bind_key__ attribute:

class User(db.Model):
    __bind_key__ = 'users'
    id = db.Column(db.Integer, primary_key=True)
    username = db.Column(db.String(80), unique=True)

Internally the bind key is stored in the table’s info dictionary as 'bind_key'. This is important to know because when you want to create a table object directly you will have to put it in there:

user_favorites = db.Table('user_favorites',
    db.Column('user_id', db.Integer, db.ForeignKey('user.id')),
    db.Column('message_id', db.Integer, db.ForeignKey('message.id')),
    info={'bind_key': 'users'}
)

If you specified the __bind_key__ on your models you can use them exactly the way you are used to. The model connects to the specified database connection itself.

Signalling Support¶

Connect to the following signals to get notified before and after changes are committed to the database. These changes are only tracked if SQLALCHEMY_TRACK_MODIFICATIONS is enabled in the config.

New in version 0.10.

Changed in version 2.1: before_models_committed is triggered correctly.

Deprecated since version 2.1: This will be disabled by default in a future version.

models_committed¶

This signal is sent when changed models were committed to the database.

The sender is the application that emitted the changes. The receiver is passed the changes parameter with a list of tuples in the form (model instance, operation).

The operation is one of 'insert', 'update', and 'delete'.

before_models_committed¶: This signal works exactly like models_committed but is emitted before the commit takes place.

Customizing¶

Flask-SQLAlchemy defines sensible defaults. However, sometimes customization is needed. There are various ways to customize how the models are defined and interacted with.

These customizations are applied at the creation of the SQLAlchemy object and extend to all models derived from its Model class.

Model Class¶

SQLAlchemy models all inherit from a declarative base class. This is exposed as db.Model in Flask-SQLAlchemy, which all models extend. This can be customized by subclassing the default and passing the custom class to model_class.

The following example gives every model an integer primary key, or a foreign key for joined-table inheritance.

Note

Integer primary keys for everything is not necessarily the best database design (that’s up to your project’s requirements), this is only an example.

from flask_sqlalchemy import Model, SQLAlchemy
import sqlalchemy as sa
from sqlalchemy.ext.declarative import declared_attr, has_inherited_table

class IdModel(Model):
    @declared_attr
    def id(cls):
        for base in cls.__mro__[1:-1]:
            if getattr(base, '__table__', None) is not None:
                type = sa.ForeignKey(base.id)
                break
        else:
            type = sa.Integer

        return sa.Column(type, primary_key=True)

db = SQLAlchemy(model_class=IdModel)

class User(db.Model):
    name = db.Column(db.String)

class Employee(User):
    title = db.Column(db.String)

Model Mixins¶

If behavior is only needed on some models rather than all models, use mixin classes to customize only those models. For example, if some models should track when they are created or updated:

from datetime import datetime

class TimestampMixin(object):
    created = db.Column(
        db.DateTime, nullable=False, default=datetime.utcnow)
    updated = db.Column(db.DateTime, onupdate=datetime.utcnow)

class Author(db.Model):
    ...

class Post(TimestampMixin, db.Model):
    ...

Query Class¶

It is also possible to customize what is available for use on the special query property of models. For example, providing a get_or method:

from flask_sqlalchemy import BaseQuery, SQLAlchemy

class GetOrQuery(BaseQuery):
    def get_or(self, ident, default=None):
        return self.get(ident) or default

db = SQLAlchemy(query_class=GetOrQuery)

# get a user by id, or return an anonymous user instance
user = User.query.get_or(user_id, anonymous_user)

And now all queries executed from the special query property on Flask-SQLAlchemy models can use the get_or method as part of their queries. All relationships defined with db.relationship (but not sqlalchemy.orm.relationship()) will also be provided with this functionality.

It also possible to define a custom query class for individual relationships as well, by providing the query_class keyword in the definition. This works with both db.relationship and sqlalchemy.relationship:

class MyModel(db.Model):
    cousin = db.relationship('OtherModel', query_class=GetOrQuery)

Note

If a query class is defined on a relationship, it will take precedence over the query class attached to its corresponding model.

It is also possible to define a specific query class for individual models by overriding the query_class class attribute on the model:

class MyModel(db.Model):
    query_class = GetOrQuery

In this case, the get_or method will be only availble on queries orginating from MyModel.query.

Model Metaclass¶

Warning

Metaclasses are an advanced topic, and you probably don’t need to customize them to achieve what you want. It is mainly documented here to show how to disable table name generation.

The model metaclass is responsible for setting up the SQLAlchemy internals when defining model subclasses. Flask-SQLAlchemy adds some extra behaviors through mixins; its default metaclass, DefaultMeta, inherits them all.

BindMetaMixin: __bind_key__ is extracted from the class and applied to the table. See Multiple Databases with Binds.
NameMetaMixin: If the model does not specify a __tablename__ but does specify a primary key, a name is automatically generated.

You can add your own behaviors by defining your own metaclass and creating the declarative base yourself. Be sure to still inherit from the mixins you want (or just inherit from the default metaclass).

Passing a declarative base class instead of a simple model base class, as shown above, to base_class will cause Flask-SQLAlchemy to use this base instead of constructing one with the default metaclass.

from flask_sqlalchemy import SQLAlchemy
from flask_sqlalchemy.model import DefaultMeta, Model

class CustomMeta(DefaultMeta):
    def __init__(cls, name, bases, d):
        # custom class setup could go here

        # be sure to call super
        super(CustomMeta, cls).__init__(name, bases, d)

    # custom class-only methods could go here

db = SQLAlchemy(model_class=declarative_base(
    cls=Model, metaclass=CustomMeta, name='Model'))

You can also pass whatever other arguments you want to declarative_base() to customize the base class as needed.

Disabling Table Name Generation¶

Some projects prefer to set each model’s __tablename__ manually rather than relying on Flask-SQLAlchemy’s detection and generation. The table name generation can be disabled by defining a custom metaclass.

from flask_sqlalchemy.model import BindMetaMixin, Model
from sqlalchemy.ext.declarative import DeclarativeMeta, declarative_base

class NoNameMeta(BindMetaMixin, DeclarativeMeta):
    pass

db = SQLAlchemy(model_class=declarative_base(
    cls=Model, metaclass=NoNameMeta, name='Model'))

This creates a base that still supports the __bind_key__ feature but does not generate table names.

API Reference¶

API¶

Configuration¶

class flask_sqlalchemy.SQLAlchemy(app=None, use_native_unicode=True, session_options=None, metadata=None, query_class=<class 'flask_sqlalchemy.BaseQuery'>, model_class=<class 'flask_sqlalchemy.model.Model'>, engine_options=None)¶

This class is used to control the SQLAlchemy integration to one or more Flask applications. Depending on how you initialize the object it is usable right away or will attach as needed to a Flask application.

There are two usage modes which work very similarly. One is binding the instance to a very specific Flask application:

app = Flask(__name__)
db = SQLAlchemy(app)

The second possibility is to create the object once and configure the application later to support it:

db = SQLAlchemy()

def create_app():
    app = Flask(__name__)
    db.init_app(app)
    return app

The difference between the two is that in the first case methods like create_all() and drop_all() will work all the time but in the second case a flask.Flask.app_context() has to exist.

By default Flask-SQLAlchemy will apply some backend-specific settings to improve your experience with them.

As of SQLAlchemy 0.6 SQLAlchemy will probe the library for native unicode support. If it detects unicode it will let the library handle that, otherwise do that itself. Sometimes this detection can fail in which case you might want to set use_native_unicode (or the SQLALCHEMY_NATIVE_UNICODE configuration key) to False. Note that the configuration key overrides the value you pass to the constructor. Direct support for use_native_unicode and SQLALCHEMY_NATIVE_UNICODE are deprecated as of v2.4 and will be removed in v3.0. engine_options and SQLALCHEMY_ENGINE_OPTIONS may be used instead.

This class also provides access to all the SQLAlchemy functions and classes from the sqlalchemy and sqlalchemy.orm modules. So you can declare models like this:

class User(db.Model):
    username = db.Column(db.String(80), unique=True)
    pw_hash = db.Column(db.String(80))

You can still use sqlalchemy and sqlalchemy.orm directly, but note that Flask-SQLAlchemy customizations are available only through an instance of this SQLAlchemy class. Query classes default to BaseQuery for db.Query, db.Model.query_class, and the default query_class for db.relationship and db.backref. If you use these interfaces through sqlalchemy and sqlalchemy.orm directly, the default query class will be that of sqlalchemy.

Check types carefully

Don’t perform type or isinstance checks against db.Table, which emulates Table behavior but is not a class. db.Table exposes the Table interface, but is a function which allows omission of metadata.

The session_options parameter, if provided, is a dict of parameters to be passed to the session constructor. See Session for the standard options.

The engine_options parameter, if provided, is a dict of parameters to be passed to create engine. See create_engine() for the standard options. The values given here will be merged with and override anything set in the 'SQLALCHEMY_ENGINE_OPTIONS' config variable or othewise set by this library.

New in version 0.10: The session_options parameter was added.

New in version 0.16: scopefunc is now accepted on session_options. It allows specifying a custom function which will define the SQLAlchemy session’s scoping.

New in version 2.1: The metadata parameter was added. This allows for setting custom naming conventions among other, non-trivial things.

The query_class parameter was added, to allow customisation of the query class, in place of the default of BaseQuery.

The model_class parameter was added, which allows a custom model class to be used in place of Model.

Changed in version 2.1: Utilise the same query class across session, Model.query and Query.

New in version 2.4: The engine_options parameter was added.

Changed in version 2.4: The use_native_unicode parameter was deprecated.

Changed in version 2.4.3: COMMIT_ON_TEARDOWN is deprecated and will be removed in version 3.1. Call db.session.commit() directly instead.

Query = None¶: Default query class used by Model.query and other queries. Customize this by passing query_class to SQLAlchemy(). Defaults to BaseQuery.

apply_driver_hacks(app, sa_url, options)¶

This method is called before engine creation and used to inject driver specific hacks into the options. The options parameter is a dictionary of keyword arguments that will then be used to call the sqlalchemy.create_engine() function.

The default implementation provides some saner defaults for things like pool sizes for MySQL and sqlite. Also it injects the setting of SQLALCHEMY_NATIVE_UNICODE.

Changed in version 2.5: Returns (sa_url, options). SQLAlchemy 1.4 made the URL immutable, so any changes to it must now be passed back up to the original caller.

apply_pool_defaults(app, options)¶: Changed in version 2.5: Returns the options dict, for consistency with apply_driver_hacks().

create_all(bind='__all__', app=None)¶: Creates all tables.

Changed in version 0.12: Parameters were added

create_engine(sa_url, engine_opts)¶

Override this method to have final say over how the SQLAlchemy engine is created.

In most cases, you will want to use 'SQLALCHEMY_ENGINE_OPTIONS' config variable or set engine_options for SQLAlchemy().

create_scoped_session(options=None)¶

Create a scoped_session on the factory from create_session().

An extra key 'scopefunc' can be set on the options dict to specify a custom scope function. If it’s not provided, Flask’s app context stack identity is used. This will ensure that sessions are created and removed with the request/response cycle, and should be fine in most cases.

Parameters: options – dict of keyword arguments passed to session class in create_session

create_session(options)¶

Create the session factory used by create_scoped_session().

The factory must return an object that SQLAlchemy recognizes as a session, or registering session events may raise an exception.

Valid factories include a Session class or a sessionmaker.

The default implementation creates a sessionmaker for SignallingSession.

Parameters: options – dict of keyword arguments passed to session class

drop_all(bind='__all__', app=None)¶: Drops all tables.

Changed in version 0.12: Parameters were added

engine¶: Gives access to the engine. If the database configuration is bound to a specific application (initialized with an application) this will always return a database connection. If however the current application is used this might raise a RuntimeError if no application is active at the moment.

get_app(reference_app=None)¶: Helper method that implements the logic to look up an application.

get_binds(app=None)¶

Returns a dictionary with a table->engine mapping.

This is suitable for use of sessionmaker(binds=db.get_binds(app)).

get_engine(app=None, bind=None)¶: Returns a specific engine.

get_tables_for_bind(bind=None)¶: Returns a list of all tables relevant for a bind.

init_app(app)¶: This callback can be used to initialize an application for the use with this database setup. Never use a database in the context of an application not initialized that way or connections will leak.

make_connector(app=None, bind=None)¶: Creates the connector for a given state and bind.

make_declarative_base(model, metadata=None)¶

Creates the declarative base that all models will inherit from.

Parameters

model – base model class (or a tuple of base classes) to pass to declarative_base(). Or a class returned from declarative_base, in which case a new base class is not created.
metadata – MetaData instance to use, or none to use SQLAlchemy’s default.

metadata¶: The metadata associated with db.Model.

reflect(bind='__all__', app=None)¶: Reflects tables from the database.

Changed in version 0.12: Parameters were added

Models¶

class flask_sqlalchemy.Model¶

Base class for SQLAlchemy declarative base model.

To define models, subclass db.Model, not this class. To customize db.Model, subclass this and pass it as model_class to SQLAlchemy.

__bind_key__¶: Optionally declares the bind to use. None refers to the default bind. For more information see Multiple Databases with Binds.

__tablename__¶: The name of the table in the database. This is required by SQLAlchemy; however, Flask-SQLAlchemy will set it automatically if a model has a primary key defined. If the __table__ or __tablename__ is set explicitly, that will be used instead.

class flask_sqlalchemy.BaseQuery(entities, session=None)¶

SQLAlchemy Query subclass with convenience methods for querying in a web application.

This is the default query object used for models, and exposed as Query. Override the query class for an individual model by subclassing this and setting query_class.

first_or_404(description=None)¶: Like first() but aborts with 404 if not found instead of returning None.

get_or_404(ident, description=None)¶: Like get() but aborts with 404 if not found instead of returning None.

paginate(page=None, per_page=None, error_out=True, max_per_page=None)¶

Returns per_page items from page page.

If page or per_page are None, they will be retrieved from the request query. If max_per_page is specified, per_page will be limited to that value. If there is no request or they aren’t in the query, they default to 1 and 20 respectively.

When error_out is True (default), the following rules will cause a 404 response:

No items are found and page is not 1.
page is less than 1, or per_page is negative.
page or per_page are not ints.

When error_out is False, page and per_page default to 1 and 20 respectively.

Returns a Pagination object.

Sessions¶

class flask_sqlalchemy.SignallingSession(db, autocommit=False, autoflush=True, **options)¶

The signalling session is the default session that Flask-SQLAlchemy uses. It extends the default session system with bind selection and modification tracking.

If you want to use a different session you can override the SQLAlchemy.create_session() function.

New in version 2.0.

New in version 2.1: The binds option was added, which allows a session to be joined to an external transaction.

get_bind(mapper=None, clause=None)¶: Return the engine or connection for a given model or table, using the __bind_key__ if it is set.

Utilities¶

class flask_sqlalchemy.Pagination(query, page, per_page, total, items)¶

Internal helper class returned by BaseQuery.paginate(). You can also construct it from any other SQLAlchemy query object if you are working with other libraries. Additionally it is possible to pass None as query object in which case the prev() and next() will no longer work.

has_next¶: True if a next page exists.

has_prev¶: True if a previous page exists

items = None¶: the items for the current page

iter_pages(left_edge=2, left_current=2, right_current=5, right_edge=2)¶

Iterates over the page numbers in the pagination. The four parameters control the thresholds how many numbers should be produced from the sides. Skipped page numbers are represented as None. This is how you could render such a pagination in the templates:

{% macro render_pagination(pagination, endpoint) %}
  <div class=pagination>
  {%- for page in pagination.iter_pages() %}
    {% if page %}
      {% if page != pagination.page %}
        <a href="{{ url_for(endpoint, page=page) }}">{{ page }}</a>
      {% else %}
        <strong>{{ page }}</strong>
      {% endif %}
    {% else %}
      <span class=ellipsis>…</span>
    {% endif %}
  {%- endfor %}
  </div>
{% endmacro %}

next(error_out=False)¶: Returns a Pagination object for the next page.

next_num¶: Number of the next page

page = None¶: the current page number (1 indexed)

pages¶: The total number of pages

per_page = None¶: the number of items to be displayed on a page.

prev(error_out=False)¶: Returns a Pagination object for the previous page.

prev_num¶: Number of the previous page.

query = None¶: the unlimited query object that was used to create this pagination object.

total = None¶: the total number of items matching the query

flask_sqlalchemy.get_debug_queries()¶

In debug mode Flask-SQLAlchemy will log all the SQL queries sent to the database. This information is available until the end of request which makes it possible to easily ensure that the SQL generated is the one expected on errors or in unittesting. If you don’t want to enable the DEBUG mode for your unittests you can also enable the query recording by setting the 'SQLALCHEMY_RECORD_QUERIES' config variable to True. This is automatically enabled if Flask is in testing mode.

The value returned will be a list of named tuples with the following attributes:

statement: The SQL statement issued
parameters: The parameters for the SQL statement
start_time / end_time: Time the query started / the results arrived. Please keep in mind that the timer function used depends on your platform. These values are only useful for sorting or comparing. They do not necessarily represent an absolute timestamp.
duration: Time the query took in seconds
context: A string giving a rough estimation of where in your application query was issued. The exact format is undefined so don’t try to reconstruct filename or function name.

Additional Information¶

License¶

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Changes¶

Version 2.5.1¶

Released 2021-03-18

Fix compatibility with Python 2.7.

Version 2.5.0¶

Released 2021-03-18

Update to support SQLAlchemy 1.4.
SQLAlchemy URL objects are immutable. Some internal methods have changed to return a new URL instead of None. #885

Version 2.4.4¶

Released 2020-07-14

Change base class of meta mixins to type. This fixes an issue caused by a regression in CPython 3.8.4. #852

Version 2.4.3¶

Released 2020-05-26

Deprecate SQLALCHEMY_COMMIT_ON_TEARDOWN as it can cause various design issues that are difficult to debug. Call db.session.commit() directly instead. #216

Version 2.4.2¶

Released 2020-05-25

Fix bad pagination when records are de-duped. #812

Version 2.4.1¶

Released 2019-09-24

Fix AttributeError when using multiple binds with polymorphic models. #651

Version 2.4.0¶

Released 2019-04-24

Make engine configuration more flexible. (#684)
Address SQLAlchemy 1.3 deprecations. (#684)
get_or_404() and first_or_404() now accept a description parameter to control the 404 message. (#636)
Use time.perf_counter for Python 3 on Windows. (#638)
Drop support for Python 2.6 and 3.3. (#687)
Add an example of Flask’s tutorial project, Flaskr, adapted for Flask-SQLAlchemy. (#720)

Version 2.3.2¶

Released 2017-10-11

Don’t mask the parent table for single-table inheritance models. (#561)

Version 2.3.1¶

Released 2017-10-05

If a model has a table name that matches an existing table in the metadata, use that table. Fixes a regression where reflected tables were not picked up by models. (#551)
Raise the correct error when a model has a table name but no primary key. (#556)
Fix repr on models that don’t have an identity because they have not been flushed yet. (#555)
Allow specifying a max_per_page limit for pagination, to avoid users specifying high values in the request args. (#542)
For paginate with error_out=False, the minimum value for page is 1 and per_page is 0. (#558)

Version 2.3.0¶

Released 2017-09-28

Multiple bugs with __tablename__ generation are fixed. Names will be generated for models that define a primary key, but not for single-table inheritance subclasses. Names will not override a declared_attr. PrimaryKeyConstraint is detected. (#541)
Passing an existing declarative_base() as model_class to SQLAlchemy.__init__ will use this as the base class instead of creating one. This allows customizing the metaclass used to construct the base. (#546)
The undocumented DeclarativeMeta internals that the extension uses for binds and table name generation have been refactored to work as mixins. Documentation is added about how to create a custom metaclass that does not do table name generation. (#546)
Model and metaclass code has been moved to a new models module. _BoundDeclarativeMeta is renamed to DefaultMeta; the old name will be removed in 3.0. (#546)
Models have a default repr that shows the model name and primary key. (#530)
Fixed a bug where using init_app would cause connectors to always use the current_app rather than the app they were created for. This caused issues when multiple apps were registered with the extension. (#547)

Version 2.2¶

Released 2017-02-27, codename Dubnium

Minimum SQLAlchemy version is 0.8 due to use of sqlalchemy.inspect.
Added support for custom query_class and model_class as args to the SQLAlchemy constructor. (#328)
Allow listening to SQLAlchemy events on db.session. (#364)
Allow __bind_key__ on abstract models. (#373)
Allow SQLALCHEMY_ECHO to be a string. (#409)
Warn when SQLALCHEMY_DATABASE_URI is not set. (#443)
Don’t let pagination generate invalid page numbers. (#460)
Drop support of Flask < 0.10. This means the db session is always tied to the app context and its teardown event. (#461)
Tablename generation logic no longer accesses class properties unless they are declared_attr. (#467)

Version 2.1¶

Released 2015-10-23, codename Caesium

Table names are automatically generated in more cases, including subclassing mixins and abstract models.
Allow using a custom MetaData object.
Add support for binds parameter to session.

Version 2.0¶

Released 2014-08-29, codename Bohrium

Changed how the builtin signals are subscribed to skip non-Flask-SQLAlchemy sessions. This will also fix the attribute error about model changes not existing.
Added a way to control how signals for model modifications are tracked.
Made the SignallingSession a public interface and added a hook for customizing session creation.
If the bind parameter is given to the signalling session it will no longer cause an error that a parameter is given twice.
Added working table reflection support.
Enabled autoflush by default.
Consider SQLALCHEMY_COMMIT_ON_TEARDOWN harmful and remove from docs.

Version 1.0¶

Released 2013-07-20, codename Aurum

Added Python 3.3 support.
Dropped 2.5 compatibility.
Various bugfixes
Changed versioning format to do major releases for each update now.

Version 0.16¶

New distribution format (flask_sqlalchemy)
Added support for Flask 0.9 specifics.

Version 0.15¶

Added session support for multiple databases.

Version 0.14¶

Make relative sqlite paths relative to the application root.

Version 0.13¶

Fixed an issue with Flask-SQLAlchemy not selecting the correct binds.

Version 0.12¶

Added support for multiple databases.
Expose BaseQuery as db.Query.
Set default query_class for db.relation, db.relationship, and db.dynamic_loader to BaseQuery.
Improved compatibility with Flask 0.7.

Version 0.11¶

Fixed a bug introduced in 0.10 with alternative table constructors.

Version 0.10¶

Added support for signals.
Table names are now automatically set from the class name unless overridden.
Model.query now always works for applications directly passed to the SQLAlchemy constructor. Furthermore the property now raises a RuntimeError instead of being None.
Added session options to constructor.
Fixed a broken __repr__.
db.Table is now a factory function that creates table objects. This makes it possible to omit the metadata.

Version 0.9¶

Applied changes to pass the Flask extension approval process.

Version 0.8¶

Added a few configuration keys for creating connections.
Automatically activate connection recycling for MySQL connections.
Added support for the Flask testing mode.

Version 0.7¶

Initial public release