Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/kayak/pypika

PyPika is a python SQL query builder that exposes the full richness of the SQL language using a syntax that reflects the resulting query. PyPika excels at all sorts of SQL queries but is especially useful for data analysis.
https://github.com/kayak/pypika

builder data functional python python3 pythonic query sql

Last synced: 4 days ago
JSON representation

PyPika is a python SQL query builder that exposes the full richness of the SQL language using a syntax that reflects the resulting query. PyPika excels at all sorts of SQL queries but is especially useful for data analysis.

Awesome Lists containing this project

README

        

PyPika - Python Query Builder
=============================

.. _intro_start:

|BuildStatus| |CoverageStatus| |Codacy| |Docs| |PyPi| |License|

Abstract
--------

What is |Brand|?

|Brand| is a Python API for building SQL queries. The motivation behind |Brand| is to provide a simple interface for
building SQL queries without limiting the flexibility of handwritten SQL. Designed with data analysis in mind, |Brand|
leverages the builder design pattern to construct queries to avoid messy string formatting and concatenation. It is also
easily extended to take full advantage of specific features of SQL database vendors.

What are the design goals for |Brand|?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

|Brand| is a fast, expressive and flexible way to replace handwritten SQL (or even ORM for the courageous souls amongst you).
Validation of SQL correctness is not an explicit goal of |Brand|. With such a large number of
SQL database vendors providing a robust validation of input data is difficult. Instead you are encouraged to check inputs you provide to |Brand| or appropriately handle errors raised from
your SQL database - just as you would have if you were writing SQL yourself.

.. _intro_end:

Read the docs: http://pypika.readthedocs.io/en/latest/

Installation
------------

.. _installation_start:

|Brand| supports python ``3.6+``. It may also work on pypy, cython, and jython, but is not being tested for these versions.

To install |Brand| run the following command:

.. code-block:: bash

pip install pypika

.. _installation_end:

Tutorial
--------

.. _tutorial_start:

The main classes in pypika are ``pypika.Query``, ``pypika.Table``, and ``pypika.Field``.

.. code-block:: python

from pypika import Query, Table, Field

Selecting Data
^^^^^^^^^^^^^^

The entry point for building queries is ``pypika.Query``. In order to select columns from a table, the table must
first be added to the query. For simple queries with only one table, tables and columns can be references using
strings. For more sophisticated queries a ``pypika.Table`` must be used.

.. code-block:: python

q = Query.from_('customers').select('id', 'fname', 'lname', 'phone')

To convert the query into raw SQL, it can be cast to a string.

.. code-block:: python

str(q)

Alternatively, you can use the `Query.get_sql()` function:

.. code-block:: python

q.get_sql()

Tables, Columns, Schemas, and Databases
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

In simple queries like the above example, columns in the "from" table can be referenced by passing string names into
the ``select`` query builder function. In more complex examples, the ``pypika.Table`` class should be used. Columns can be
referenced as attributes on instances of ``pypika.Table``.

.. code-block:: python

from pypika import Table, Query

customers = Table('customers')
q = Query.from_(customers).select(customers.id, customers.fname, customers.lname, customers.phone)

Both of the above examples result in the following SQL:

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers

An alias for the table can be given using the ``.as_`` function on ``pypika.Table``

.. code-block:: sql

customers = Table('x_view_customers').as_('customers')
q = Query.from_(customers).select(customers.id, customers.phone)

.. code-block:: sql

SELECT id,phone FROM x_view_customers customers

A schema can also be specified. Tables can be referenced as attributes on the schema.

.. code-block:: sql

from pypika import Table, Query, Schema

views = Schema('views')
q = Query.from_(views.customers).select(customers.id, customers.phone)

.. code-block:: sql

SELECT id,phone FROM views.customers

Also references to databases can be used. Schemas can be referenced as attributes on the database.

.. code-block:: sql

from pypika import Table, Query, Database

my_db = Database('my_db')
q = Query.from_(my_db.analytics.customers).select(customers.id, customers.phone)

.. code-block:: sql

SELECT id,phone FROM my_db.analytics.customers

Results can be ordered by using the following syntax:

.. code-block:: python

from pypika import Order
Query.from_('customers').select('id', 'fname', 'lname', 'phone').orderby('id', order=Order.desc)

This results in the following SQL:

.. code-block:: sql

SELECT "id","fname","lname","phone" FROM "customers" ORDER BY "id" DESC

Arithmetic
""""""""""

Arithmetic expressions can also be constructed using pypika. Operators such as `+`, `-`, `*`, and `/` are implemented
by ``pypika.Field`` which can be used simply with a ``pypika.Table`` or directly.

.. code-block:: python

from pypika import Field

q = Query.from_('account').select(
Field('revenue') - Field('cost')
)

.. code-block:: sql

SELECT revenue-cost FROM accounts

Using ``pypika.Table``

.. code-block:: python

accounts = Table('accounts')
q = Query.from_(accounts).select(
accounts.revenue - accounts.cost
)

.. code-block:: sql

SELECT revenue-cost FROM accounts

An alias can also be used for fields and expressions.

.. code-block:: sql

q = Query.from_(accounts).select(
(accounts.revenue - accounts.cost).as_('profit')
)

.. code-block:: sql

SELECT revenue-cost profit FROM accounts

More arithmetic examples

.. code-block:: python

table = Table('table')
q = Query.from_(table).select(
table.foo + table.bar,
table.foo - table.bar,
table.foo * table.bar,
table.foo / table.bar,
(table.foo+table.bar) / table.fiz,
)

.. code-block:: sql

SELECT foo+bar,foo-bar,foo*bar,foo/bar,(foo+bar)/fiz FROM table

Filtering
"""""""""

Queries can be filtered with ``pypika.Criterion`` by using equality or inequality operators

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id, customers.fname, customers.lname, customers.phone
).where(
customers.lname == 'Mustermann'
)

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers WHERE lname='Mustermann'

Query methods such as select, where, groupby, and orderby can be called multiple times. Multiple calls to the where
method will add additional conditions as

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id, customers.fname, customers.lname, customers.phone
).where(
customers.fname == 'Max'
).where(
customers.lname == 'Mustermann'
)

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers WHERE fname='Max' AND lname='Mustermann'

Filters such as IN and BETWEEN are also supported

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id,customers.fname
).where(
customers.age[18:65] & customers.status.isin(['new', 'active'])
)

.. code-block:: sql

SELECT id,fname FROM customers WHERE age BETWEEN 18 AND 65 AND status IN ('new','active')

Filtering with complex criteria can be created using boolean symbols ``&``, ``|``, and ``^``.

AND

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id, customers.fname, customers.lname, customers.phone
).where(
(customers.age >= 18) & (customers.lname == 'Mustermann')
)

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers WHERE age>=18 AND lname='Mustermann'

OR

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id, customers.fname, customers.lname, customers.phone
).where(
(customers.age >= 18) | (customers.lname == 'Mustermann')
)

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers WHERE age>=18 OR lname='Mustermann'

XOR

.. code-block:: python

customers = Table('customers')
q = Query.from_(customers).select(
customers.id, customers.fname, customers.lname, customers.phone
).where(
(customers.age >= 18) ^ customers.is_registered
)

.. code-block:: sql

SELECT id,fname,lname,phone FROM customers WHERE age>=18 XOR is_registered

Convenience Methods
"""""""""""""""""""

In the `Criterion` class, there are the static methods `any` and `all` that allow building chains AND and OR expressions with a list of terms.

.. code-block:: python

from pypika import Criterion

customers = Table('customers')
q = Query.from_(customers).select(
customers.id,
customers.fname
).where(
Criterion.all([
customers.is_registered,
customers.age >= 18,
customers.lname == "Jones",
])
)

.. code-block:: sql

SELECT id,fname FROM customers WHERE is_registered AND age>=18 AND lname = "Jones"

Grouping and Aggregating
""""""""""""""""""""""""

Grouping allows for aggregated results and works similar to ``SELECT`` clauses.

.. code-block:: python

from pypika import functions as fn

customers = Table('customers')
q = Query \
.from_(customers) \
.where(customers.age >= 18) \
.groupby(customers.id) \
.select(customers.id, fn.Sum(customers.revenue))

.. code-block:: sql

SELECT id,SUM("revenue") FROM "customers" WHERE "age">=18 GROUP BY "id"

After adding a ``GROUP BY`` clause to a query, the ``HAVING`` clause becomes available. The method
``Query.having()`` takes a ``Criterion`` parameter similar to the method ``Query.where()``.

.. code-block:: python

from pypika import functions as fn

payments = Table('payments')
q = Query \
.from_(payments) \
.where(payments.transacted[date(2015, 1, 1):date(2016, 1, 1)]) \
.groupby(payments.customer_id) \
.having(fn.Sum(payments.total) >= 1000) \
.select(payments.customer_id, fn.Sum(payments.total))

.. code-block:: sql

SELECT customer_id,SUM(total) FROM payments
WHERE transacted BETWEEN '2015-01-01' AND '2016-01-01'
GROUP BY customer_id HAVING SUM(total)>=1000

Joining Tables and Subqueries
"""""""""""""""""""""""""""""

Tables and subqueries can be joined to any query using the ``Query.join()`` method. Joins can be performed with either
a ``USING`` or ``ON`` clauses. The ``USING`` clause can be used when both tables/subqueries contain the same field and
the ``ON`` clause can be used with a criterion. To perform a join, ``...join()`` can be chained but then must be
followed immediately by ``...on()`` or ``...using(*field)``.

Join Types
~~~~~~~~~~

All join types are supported by |Brand|.

.. code-block:: python

Query \
.from_(base_table)
...
.join(join_table, JoinType.left)
...

.. code-block:: python

Query \
.from_(base_table)
...
.left_join(join_table) \
.left_outer_join(join_table) \
.right_join(join_table) \
.right_outer_join(join_table) \
.inner_join(join_table) \
.outer_join(join_table) \
.full_outer_join(join_table) \
.cross_join(join_table) \
.hash_join(join_table) \
...

See the list of join types here ``pypika.enums.JoinTypes``

Example of a join using `ON`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: python

history, customers = Tables('history', 'customers')
q = Query \
.from_(history) \
.join(customers) \
.on(history.customer_id == customers.id) \
.select(history.star) \
.where(customers.id == 5)

.. code-block:: sql

SELECT "history".* FROM "history" JOIN "customers" ON "history"."customer_id"="customers"."id" WHERE "customers"."id"=5

As a shortcut, the ``Query.join().on_field()`` function is provided for joining the (first) table in the ``FROM`` clause
with the joined table when the field name(s) are the same in both tables.

Example of a join using `ON`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: python

history, customers = Tables('history', 'customers')
q = Query \
.from_(history) \
.join(customers) \
.on_field('customer_id', 'group') \
.select(history.star) \
.where(customers.group == 'A')

.. code-block:: sql

SELECT "history".* FROM "history" JOIN "customers" ON "history"."customer_id"="customers"."customer_id" AND "history"."group"="customers"."group" WHERE "customers"."group"='A'

Example of a join using `USING`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: python

history, customers = Tables('history', 'customers')
q = Query \
.from_(history) \
.join(customers) \
.using('customer_id') \
.select(history.star) \
.where(customers.id == 5)

.. code-block:: sql

SELECT "history".* FROM "history" JOIN "customers" USING "customer_id" WHERE "customers"."id"=5

Example of a correlated subquery in the `SELECT`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: python

history, customers = Tables('history', 'customers')
last_purchase_at = Query.from_(history).select(
history.purchase_at
).where(history.customer_id==customers.customer_id).orderby(
history.purchase_at, order=Order.desc
).limit(1)
q = Query.from_(customers).select(
customers.id, last_purchase_at.as_('last_purchase_at')
)

.. code-block:: sql

SELECT
"id",
(SELECT "history"."purchase_at"
FROM "history"
WHERE "history"."customer_id" = "customers"."customer_id"
ORDER BY "history"."purchase_at" DESC
LIMIT 1) "last_purchase_at"
FROM "customers"

Unions
""""""

Both ``UNION`` and ``UNION ALL`` are supported. ``UNION DISTINCT`` is synonymous with ``UNION`` so |Brand| does not
provide a separate function for it. Unions require that queries have the same number of ``SELECT`` clauses so
trying to cast a unioned query to string will throw a ``SetOperationException`` if the column sizes are mismatched.

To create a union query, use either the ``Query.union()`` method or `+` operator with two query instances. For a
union all, use ``Query.union_all()`` or the `*` operator.

.. code-block:: python

provider_a, provider_b = Tables('provider_a', 'provider_b')
q = Query.from_(provider_a).select(
provider_a.created_time, provider_a.foo, provider_a.bar
) + Query.from_(provider_b).select(
provider_b.created_time, provider_b.fiz, provider_b.buz
)

.. code-block:: sql

SELECT "created_time","foo","bar" FROM "provider_a" UNION SELECT "created_time","fiz","buz" FROM "provider_b"

Intersect
"""""""""

``INTERSECT`` is supported. Intersects require that queries have the same number of ``SELECT`` clauses so
trying to cast a intersected query to string will throw a ``SetOperationException`` if the column sizes are mismatched.

To create a intersect query, use the ``Query.intersect()`` method.

.. code-block:: python

provider_a, provider_b = Tables('provider_a', 'provider_b')
q = Query.from_(provider_a).select(
provider_a.created_time, provider_a.foo, provider_a.bar
)
r = Query.from_(provider_b).select(
provider_b.created_time, provider_b.fiz, provider_b.buz
)
intersected_query = q.intersect(r)

.. code-block:: sql

SELECT "created_time","foo","bar" FROM "provider_a" INTERSECT SELECT "created_time","fiz","buz" FROM "provider_b"

Minus
"""""

``MINUS`` is supported. Minus require that queries have the same number of ``SELECT`` clauses so
trying to cast a minus query to string will throw a ``SetOperationException`` if the column sizes are mismatched.

To create a minus query, use either the ``Query.minus()`` method or `-` operator with two query instances.

.. code-block:: python

provider_a, provider_b = Tables('provider_a', 'provider_b')
q = Query.from_(provider_a).select(
provider_a.created_time, provider_a.foo, provider_a.bar
)
r = Query.from_(provider_b).select(
provider_b.created_time, provider_b.fiz, provider_b.buz
)
minus_query = q.minus(r)

(or)

minus_query = Query.from_(provider_a).select(
provider_a.created_time, provider_a.foo, provider_a.bar
) - Query.from_(provider_b).select(
provider_b.created_time, provider_b.fiz, provider_b.buz
)

.. code-block:: sql

SELECT "created_time","foo","bar" FROM "provider_a" MINUS SELECT "created_time","fiz","buz" FROM "provider_b"

EXCEPT
""""""

``EXCEPT`` is supported. Minus require that queries have the same number of ``SELECT`` clauses so
trying to cast a except query to string will throw a ``SetOperationException`` if the column sizes are mismatched.

To create a except query, use the ``Query.except_of()`` method.

.. code-block:: python

provider_a, provider_b = Tables('provider_a', 'provider_b')
q = Query.from_(provider_a).select(
provider_a.created_time, provider_a.foo, provider_a.bar
)
r = Query.from_(provider_b).select(
provider_b.created_time, provider_b.fiz, provider_b.buz
)
minus_query = q.except_of(r)

.. code-block:: sql

SELECT "created_time","foo","bar" FROM "provider_a" EXCEPT SELECT "created_time","fiz","buz" FROM "provider_b"

Date, Time, and Intervals
"""""""""""""""""""""""""

Using ``pypika.Interval``, queries can be constructed with date arithmetic. Any combination of intervals can be
used except for weeks and quarters, which must be used separately and will ignore any other values if selected.

.. code-block:: python

from pypika import functions as fn

fruits = Tables('fruits')
q = Query.from_(fruits) \
.select(fruits.id, fruits.name) \
.where(fruits.harvest_date + Interval(months=1) < fn.Now())

.. code-block:: sql

SELECT id,name FROM fruits WHERE harvest_date+INTERVAL 1 MONTH QueryBuilder:
if isinstance(col, str):
col = Field(col)

return query.where(col > fn.Now() - num_days)

def count_groups(query: QueryBuilder, *groups) -> QueryBuilder:
return query.groupby(*groups).select(*groups, fn.Count("*").as_("n_rows"))

base_query = Query.from_("table")

query = (
base_query
.pipe(filter_days, "date", num_days=7)
.pipe(count_groups, "col1", "col2")
)

This produces:

.. code-block:: sql

SELECT "col1","col2",COUNT(*) n_rows
FROM "table"
WHERE "date">NOW()-7
GROUP BY "col1","col2"

.. _tutorial_end:

.. _contributing_start:

Contributing
------------

We welcome community contributions to |Brand|. Please see the `contributing guide <6_contributing.html>`_ to more info.

.. _contributing_end:

.. _license_start:

License
-------

Copyright 2020 KAYAK Germany, GmbH

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

Crafted with ♥ in Berlin.

.. _license_end:

.. _appendix_start:

.. |Brand| replace:: *PyPika*

.. _appendix_end:

.. _available_badges_start:

.. |BuildStatus| image:: https://github.com/kayak/pypika/workflows/Unit%20Tests/badge.svg
:target: https://github.com/kayak/pypika/actions
.. |CoverageStatus| image:: https://coveralls.io/repos/kayak/pypika/badge.svg?branch=master
:target: https://coveralls.io/github/kayak/pypika?branch=master
.. |Codacy| image:: https://api.codacy.com/project/badge/Grade/6d7e44e5628b4839a23da0bd82eaafcf
:target: https://www.codacy.com/app/twheys/pypika
.. |Docs| image:: https://readthedocs.org/projects/pypika/badge/?version=latest
:target: http://pypika.readthedocs.io/en/latest/
.. |PyPi| image:: https://img.shields.io/pypi/v/pypika.svg?style=flat
:target: https://pypi.python.org/pypi/pypika
.. |License| image:: https://img.shields.io/hexpm/l/plug.svg?maxAge=2592000
:target: http://www.apache.org/licenses/LICENSE-2.0

.. _available_badges_end: