dsql

Dead simple RDBMS handling library

https://github.com/gwn/dsql

Dead simple SQL generation and result handling library. Designed to work with Python DB API 2 compatible database connection objects.

Install with::

pip install dsql

You can use dsql in two ways. First case is SQL statement generation::

>>> from dsql import buildquery

>>> buildquery('select',
               'people',
               ['name', 'surname'],
               where=[{'age >': 30}],
               orderby='-age',
               dialect='postgresql')

(
    'SELECT "name", "surname" FROM "people" WHERE "age" > %s ORDER BY "age" DESC',
    [30]
)

Second use case is to create a manager object that, in addition to generating your statements, automatically executes them and handles the results for you::

>>> import psycopg2
>>> from psycopg2.extras import DictCursor
>>> from dsql import makemanager

>>> conn = psycopg2.connect(database='foo', cursor_factory=DictCursor)

>>> db = dsql.makemanager(conn, dialect='postgresql')

>>> itemiter = db.select('products', where=[{'color =': 'red'}])

>>> itemiter.next()
{
    'id': 1,
    'title': 'Shirt',
    'color': 'red'
}

>>> db.insert('products', [{'title': 'Pants', 'color': 'green'},
>>>                        {'title': 'Socks', 'color': 'yellow'}])
[2, 3]

Note that it is required to configure the connection to return DictCursors instead of standard cursors, as in the example above.

Check out the reference section below for the documentation of the whole API.

Installation::

pip install dsql

Reference

Check out::

# Query Builder

query, params = dsql.buildquery(operation, tablename,
                                <depends-on-the-operation>, ...
                                dialect='standard')

query, params = dsql.buildquery('select', tablename, fieldlist=[],
                                where=[], groupby=[], having=[],
                                orderby=[], limit=0, offset=0,
                                dialect='standard')

query, params = dsql.buildquery('insert', tablename, recordlist,
                                dialect='standard')

query, params = dsql.buildquery('update', tablename, updates, where=[],
                                orderby=[], limit=0, offset=0,
                                dialect='standard')

query, params = dsql.buildquery('delete', tablename, where=[], orderby=[],
                                dialect='standard')

query, params = dsql.buildquery('raw', query, params)


# Manager

db = dsql.makemanager(dbapi2_compatible_connection_object,
                      dialect='standard')

itemiter = db.select(tablename, fieldlist=[], where=[], groupby=[],
                     having=[], orderby=[], limit=1, offset=0, commit=True,
                     dry_run=False, response_handler=None)

list_of_inserted_ids = db.insert(tablename, recordlist, commit=True,
                                 dry_run=False, response_handler=None)

number_of_affected_rows = db.update(tablename, updates, where=[],
                                    orderby=[], limit=0, offset=0,
                                    commit=True, dry_run=False,
                                    response_handler=None)

number_of_affected_rows = db.delete(tablename, where=[], orderby=[],
                                    commit=True, dry_run=False,
                                    response_handler=None)

mixed = db.raw(query, params, commit=True, dry_run=False,
               response_handler=None)
# return value of this one depends on the type of query.

related_connection_object = db.conn

Documentation of common parameters:

fieldlist

List of fields, such as ['name', 'age', 'occupation']. Pass an empty list, or skip altogether, to get all the fields.

where

List of condition groups.

Each condition group is a dict of predicate and value pairs, such as: {'name =': 'John', 'age >': 30}. Each pair is combined with AND, so this example is translated to the template "name" = %s AND "age" > %s and values of ['John', 30].

Condition groups themselves are combined with OR, so the following where expression::

[{'name =': 'John', 'age >': 30}, {'occupation in': ['engineer', 'artist']}]

Translates to::

WHERE ("name" = %s AND age > %s) OR (occupation IN (%s, %s))

with the values of: ['John', 30, 'engineer', 'artist']

All standard comparison operators along with LIKE, NOT LIKE, IN and NOT IN are supported.

If you need to construct more complicated filters, try raw queries.

groupby

List of group fields, such as ['age', 'occupation']

having

Same as where.

orderby

List of fields to order by. Add the - prefix to field names for descending order. Example: ['age', '-net_worth']

limit

Limit as an integer, such as 50.

offset

Offset as an integer, such as 200.

dialect

standard, postgresql or mysql.

commit

Automatically commit the query. If you choose not to commit, you can always get the connection object from the manager object (via manager.conn) and make the commit yourself when the time is right.

dry_run

True or False. If True, does not execute the query, but dump it to the standard error for inspecting.

response_handler

By default, the manager object handles the responses for you. It returns an iterator of records for select calls, list of last inserted ids for insert calls, and number of affected rows for others. In the cases you want to handle the response yourself, you can pass your own response_handler whose arguments will be the cursor object and the current dialect. Example::

value_of_custom_handler = manager.select(tablename, limit=10,
                                         response_handler=custom_handler)

Examples

PosgreSQL with psycopg2::

import psycopg2
import psycopg2.extras
import dsql

conn = psycopg2.connect(host='localhost', user='root', database='lorem',
                        cursor_factory=psycopg2.extras.DictCursor)

db = dsql.makemanager(conn, dialect='postgresql')

itemiter = db.select('products', ['id', 'name', 'description'])
item = itemiter.next()
print item['name']

...

MySQL with MySQLdb::

import MySQLdb
import MySQLdb.cursors
import dsql

conn = MySQLdb.connect(host='localhost', user='root', db='lorem',
                       cursorclass=MySQLdb.cursors.DictCursor)

db = dsql.makemanager(conn, dialect='mysql')

itemiter = db.select('products',
                     ['id', 'name', 'description'],
                     where=[{'status =': 'in stock'}])
item = itemiter.next()
print item['name']

last_insert_ids = db.insert('products',
                            [
                                {
                                   'name': 'foo',
                                   'description': 'what a product!',
                                }
                            ])

last_insert_ids = db.insert('products',
                            [
                               {
                                   'name': 'foo',
                                   'description': 'what a product!',
                               }
                            ],
                            commit=False)
db.conn.commit()

affected_rowcount = db.update('products',
                              {'name': 'lorem ipsum'},
                              where=[{'id =': 888}])

affected_rowcount = db.delete('products', where=[{'id =': 777}])