Python-SQL Abstraction Layer

When developing a (Web) application, you end up designing a layer, where 90% of the SQL queries are very simple and the rest are really complex and ugly, some even dynamically generated ones. This package provides following capabilities:

PostgreSQL support
The module was tested with pyPgSQL. If you tried it with another module and it does not work, tell me about it and i will try to fix it.
Dynamic configuration on startup - reads tables, views, functions from DB, optionally reads even table columns to provide better warning when trying to access non-existent columns. If you have schema with many thousand tables, you can disable even the basic schema loading, so that the python code does not hold all the table instances in memory.
Easy queries look simple
Hard queries look somewhat better than raw SQL, but in my opinion raw SQL is still better understandable.
Dynamically generated queries are EASY.
Table/view inserts,select, update
Calling of Postgres stored procedures
inner/outer joins, union, intersect, except
Subqueries (in the WHERE and FROM clause)
'Case' generated columns (automatic translation from python code)
Automatic grouping operations
Heavy operator overloading, so that the majority of operations can be written directly in python code.
Simple expression can be generated in one-line code - important for web template languages
Can generate SQL query with placeholders (%(name)s syntax)

The module, however, does not free you from learning SQL first. Althouth the module does a lot of error checking, it is possible to generate incorrect SQL queries. If you do not know SQL, you will have problems debugging the code.

Installation

Download sqlabstr-0.5.tar.gz, extract the files and install it to python site-packages directory.

# python setup.py install