OVERVIEW: * clean API * flexible usage * good performance * 100% of unit tests pass * 100% code coverage (currently, only 75 lines of code) * each wrapper function can optionally have straight sql passed to it * wrapper functions use the same ordering of arguments as sql does * so it is easy for those who know sql to use the wrappers (thus saving keystrokes!) * runs on any platform python2.5 runs (it would be trivial to port to earlier versions of python)

Each sql command can be run in one line of code.

The results are returned in lists of lightweight objects (db_row.py) which have an optional implementation in C that is used if installed. The row objects give many methods of accessing the data like:

row.fields.fieldname row['fieldname'] row[1] # complete backwards compatibility with existing DBAPI return values

USAGE: ```

import sqlite3 # or any DBAPI compliant database module import edendb conn = sqlite3.connect(filename) db = edendb.EdenDB(conn)

data manipulation statments

db.dml("create table people (name varchar(10), age integer, pet varchar(10))")

insert

cols = ['name','age','pet'] db.insert('people', cols, ['paul',102,'cat']) db.insert('people', cols, ['tom',88,'dog']) db.insert('people', cols, ['janise',10,'butterfly']) db.insert('people', cols, ['guido',2,'snake'])

delete

db.delete('people', 'age = 10')

selecting

db.select_one_value('count(*) as cnt', 'people') 3 rows = db.select(['name','age'], 'people', 'age > 30', iter=True) for row in rows: # rows is an iterator so the database rows are loaded into memory one at a time print row.fields.field1, row['field2'], row[0]

without the iter=True in the select method a list of rows is returned

db.select(['name','age'], 'people', 'age > 30') [(u'paul', 102), (u'tom', 88)]

update

db.update('people', ['name'], ['Paul'], "name = 'paul'") row = db.select_one_row(['*'], 'people', "name = 'Paul'") row (u'Paul', 102, u'cat') row[1] 102 row.fields.pet u'cat' row['name'] u'Paul' conn.close() ```

See the doctests in eden_db.py for more examples.

TESTS: cd edendb python test.py -h

LIMITATIONS:

• All of those listed in db_row.py
• Using sql functions that do not return from the db driver with column names (like count(*) (at least on my testing in sqlite)) does not work. You have to use an alias. (a limitation of db_row.py) For example. select count(*) from tablename # this throws an exeption select count(*) as cnt from tablename # this works

db_row.py AND C IMPLEMENTATION:

To get the speedup and low memory overhead of the C implementation, download and install the following (assuming you are on Ubuntu).

sudo aptitude install build-essentials python-dev wget http://opensource.theopalgroup.com/files/db_row-0.8.tgz tar -zxvf db_row-0.8.tgz cd db_row-0.8 sudo python setup.py install

Incidentally, the db_row.py file is the same as the one from the above tarball.

Thanks to Kevin Jacobs (jacobs@theopalgroup.com) for the awesome db_row module!