sheba 0.0.2

Motivation

I can write SQL faster than I can figure out the various library and ORM abstraction layers. Writing SQL directly in code is pretty disgusting. So I stole an idea from the Axamol SQL Library and along with a couple updates have created this library.

What is it?

First, we’ll consider some basic query definitions:

name: create_roles_table
type: update
sql: |
    CREATE  TABLE roles
    (
        scene   text,
        name    text,
        actor   text,
        UNIQUE(scene, name)
    )
---
name: add_role
type: update
sql: |
    INSERT
    INTO    roles(scene, name, actor)
    VALUES  (${scene}, ${name}, ${actor})
---
name: list_roles
sql: |
    SELECT  name,
            actor
    FROM    roles
    ORDER   BY
            name ASC

This is a YAML file that defines a couple queries for working with a simple table that lists scenes, roles, and actors. This is the contents of the ./test/yaml/conn-queries.yaml that’s used in a couple of the test cases in case you’re curious.

Now, to use these in some code:

>>> import sheba
>>> lib = sheba.Library.from_file('test/yaml/conn-queries.yaml')
>>> conn = sheba.connect(lib, 'sqlite3', ':memory:')

Now that we have a connection, lets create the roles table:

>>> conn.u.create_roles_table()
-1

The -1 is due to DDL statements not returning row information. I could technically create a third class of statement types so avoid this, but in the not doctest world you can just ignore that return value.

Next we’ll insert a couple rows:

>>> conn.u.add_role(scene="Parrot Sketch", name="MR PRALINE", actor="John Cleese")
1
>>> conn.u.add_role(scene="Parrot Sketch", name="SHOP OWNER", actor="Machale Palin")
1
>>> conn.u.add_role(scene="Parrot Sketch", name="DEAD PARROT", actor="Fake Parrot")
1

Yep, it’s that simple. And lastly, we’ll list the roles in the table:

>>> for row in conn.q.list_roles():
...     print "Name: %(name)s Actor: %(actor)s" % row
...
Name: DEAD PARROT Actor: Fake Parrot
Name: MR PRALINE Actor: John Cleese
Name: SHOP OWNER Actor: Machale Palin

As they say, “Wicked awesome.”

Possible Query Attributes

The example above only used name, type, and sql. Only name and sql are absolutely required. The full list of attributes is:

• name - The name used to reference this query from Python

• desc - An explanation of the query for documentation.

• type - query or update. Defaults to query.

• dbs - A list of database names that this SQL will work with. The default

  value is None which is interpreted as “Use this query when no SQL has been defined for the current connections database connection.”

• sql - The actual query. SQL is passed through Mako with access to

  any parameter names that were provide to the query. To bind a parameter in a query, simply print the value with standard Mako syntax like ${my_parameter_name}. Sheba will automatically replace it with the proper bind variable syntax and pass the supplied value onto the database connection.

License

Released under the MIT license. See the LICENSE file for more details.