pgwrap 0.7

Simple PostgreSQL database wrapper - provides wrapper over psycopg2 supporting a Python API for common sql functions

pgwrap - simple PostgreSQL database wrapper

The 'pgwrap' module provides a simple wrapper over psycopg2 supporting a Python API for common sql functions.

This is not intended to provide ORM-like functionality, just to make it easier to interact with PostgreSQL from python code for simple use-cases and allow direct SQL access for more complex operations.

The module wraps the excellent 'psycopg2' library and most of the functionality is provided by this behind the scenes.

The module provides:

• Simplified handling of connections/cursor

  • Connection pool (provided by psycopg2.pool)
  • Cursor context handler

• Python API to wrap basic SQL functionality

  • Simple select,update,delete,join methods extending the cursor context handler (also available as stand-alone methods which create an implicit cursor for simple queries)

• Query results as dict (using psycopg2.extras.DictCursor)

• Callable prepared statements

• Logging support

Basic usage

>>> import pgwrap
>>> db = pgwrap.connection(url='postgres://localhost')
>>> with db.cursor() as c:
...     c.query('select version()')
[['PostgreSQL...']]
>>> v = db.query_one('select version()')
>>> v
['PostgreSQL...']
>>> v.items()
[('version', 'PostgreSQL...')]
>>> v['version']
'PostgreSQL...'

Connection

The connection class initialises an internal connection pool and provides methods to return a cursor object or execute SQL queries directly (using an implicit cursor).

The intention is that a single instance of this class is created at application start up.

Cursor

The module provides a cursor context handler wrapping the psycopg2 cursor.

Entering the cursor context handler will obtain a connection from the connection pool and create a cursor using this connection. When the context handler is exited the associated transaction will be committed, cursor closed, and the connection released back to the connection pool.

The cursor object uses the psycopg2 'DictCursor' by default (which returns rows as a pseudo python dictionary) however this can be overridden by providing a 'cursor_factory' parameter to the constructor.

>>> db = pgwrap.connection(url='postgres://localhost')
>>> with db.cursor() as c:
...     c.query('select version()')
[['PostgreSQL...']]

The cursor context provides the following basic methods:

execute - execute SQL query and return rowcount query - execute SQL query and fetch results query_one - execute SQL query and fetch first result query_dict - execute SQL query and return results as dict

System Message: ERROR/3 (<string>, line 78)

Unexpected indentation.

keyed on specified key (which should be unique)

System Message: WARNING/2 (<string>, line 79)

Block quote ends without a blank line; unexpected unindent.

commit - Commit transaction (called implicitly on exiting

context handler)

System Message: WARNING/2 (<string>, line 81)

Definition list ends without a blank line; unexpected unindent.

rollback - Rollback transaction

In addition the cursor can use the SQL API methods described below or access the underlying psycopg2 cursor (via the self.cursor attribute).

The cursor methods are also available as standalone functions which run inside an implicit cursor object.

SQL API

The cursor class also provides a simple Python API for common SQL operations. The basic methods provides are:

select - single table select (with corresponding select_one,

select_dict methods)

join - two table join (with corresponding join_one,

join_dict methods)

System Message: WARNING/2 (<string>, line 99)

Definition list ends without a blank line; unexpected unindent.

insert - SQL insert update - SQL update delete - SQL delete

The methods can be parameterised to customise the associated query (see db module for detail):

where - 'where' clause as dict (column operators can be

specified using the colunm__operator format)

where = {'name':'abc','status__in':(1,2,3)}

columns - list of columns to be returned - these can

be real columns or expressions. If spefified as a tuple the column is explicitly named using the AS operator

columns = ('name',('status > 1','updated'))

order - sort order as list (use 'column__desc' to

reverse order)

order = ('name__desc',)

limit - row limit (int)

offset - offset (int)

on - join columns (as tuple)

values - insert data as dict

returning - columns to return (string)

The methods are also available as standalone functions which create an implicit cursor object.

Basic usage:

>>> db.create_table('t1','id serial,name text,count int')
>>> db.create_table('t2','id serial,t1_id int,value text')
>>> db.log = sys.stdout
>>> db.insert('t1',{'name':'abc','count':0},returning='id,name')
INSERT INTO t1 (name) VALUES ('abc') RETURNING id,name
[1, 'abc']
>>> db.insert('t2',{'t1_id':1,'value':'t2'})
INSERT INTO t2 (t1_id,value) VALUES (1,'t2')
1
>>> db.select('t1')
SELECT * FROM t1
[[1, 'abc', 0]]
>>> db.select_one('t1',where={'name':'abc'},columns=('name','count'))
SELECT name, count FROM t1 WHERE name = 'abc'
['abc', 0]
>>> db.join(('t1','t2'),columns=('t1.id','t2.value'))
SELECT t1.id, t2.value FROM t1 JOIN t2 ON t1.id = t2.t1_id
[[1, 't2']]
>>> db.insert('t1',{'name':'abc'},returning='id')
INSERT INTO t1 (name) VALUES ('abc') RETURNING id
[2]
>>> db.update('t1',{'name':'xyz'},where={'name':'abc'})
UPDATE t1 SET name = 'xyz' WHERE name = 'abc'
2
>>> db.update('t1',{'count__func':'count + 1'},where={'count__lt':10},returning="id,count")
UPDATE t1 SET count = count + 1 WHERE count < 10 RETURNING id,count
[[1, 1]]

Prepared Statements

Prepared statements can be created using the

connection.prepare(stmt,params,name,call_type)

stmt : prepared statement (with parameters identified

in the statement using the psql $1,$2... notation)

params : list of optional parameter types (usually not

needed - infered by psql)

name : name for the prepared statement (usually

autogenerated)

call_type : method used when instance called as method

(defaults to 'query')

The constructor returns a PreparedStatement object which can be used instead of an sql statement in the connection.execute and connection.query_xxx methods.

>>> p = db.prepare('UPDATE t1 SET name = $2 WHERE id = $1')
PREPARE _pstmt_001  AS UPDATE t1 SET name = $2 WHERE id = $1
>>> with db.cursor() as c:
...     c.execute(p,(1,'xxx'))
EXECUTE _pstmt_001 (1,'xxx')

The PreparedStatement object can also be called directly using the execute/query/query_one/query_dict methods. The instance is also directly callable using the method type identified in 'call_type'

>>> p = db.prepare('UPDATE t1 SET name = $2 WHERE id = $1')
PREPARE _pstmt_001  AS UPDATE t1 SET name = $2 WHERE id = $1
>>> p.execute(1,'xxx')
EXECUTE _pstmt_001 (1,'xxx')
>>> p(1,'xxx')
EXECUTE _pstmt_001 (1,'xxx')

Logging

To enable logging the connection.log attribute can be set to either an instance of logging.Logger or a file-like object (supporting the write method).

The log message is generated using the self.logf function (called with the cursor object as a parameter). By default this just returns the query string however can be customised as needed. A cursor.timestamp attribute is available to allow execution time to be tracked.

>>> db.log = sys.stdout
>>> db.logf = lambda c : '[%f] %s' % (time.time() - c.timestamp,c.query)
>>> db.query('SELECT * FROM t1')
[0.000536] SELECT * FROM t1

Changelog

• 0.1 19-10-2012 Initial import
• 0.2 20-10-2012 Remove psycopg2 dep in setup.py
• 0.3 20-10-2012 Remove hstore default for cursor
• 0.4 21-10-2012 Add logging support
• 0.5 22-12-2012 Refactor connection class / remove globals
• 0.6 23-12-2012 Add support for prepared statements
• 0.7 26-12-2012 Add callable prepared statements & named cursor

Author

• Paul Chakravarti (paul.chakravarti@gmail.com)

Master Repository/Issues

• https://github.com/paulchakravarti/pgwrap