pysqlite multiple connections

[Gerhard Häring]

* Tertius Cronje <tcronj at yahoo.co.uk> [2003-05-08 00:24 +0200]:
> Gerhard, thank you so much.
> 
> This gives me a few more options to consider.  The application I want
> to use it for will probably not be used by more than 10 users. I'll
> have to build some extra logic into the code to enable single use of a
> cursor.

You can have multiple cursors per connection, that's no problem.
Transactions are started implicitely by opening the first cursor for a
connection. Then you use .commit() and .rollback() on the connection
object. Unless you turn off transactions by setting autocommit=True, of
course.

> I was looking for some simple code examples but without any luck.

My small apps usually look like this:

There is a module for the database stuff that looks approx. like:

#v+
#  Module db.py
con = None

def open_con():
    con = sqlite.connect(...)

def getcon():
    return con
#v-

Now somewhere in your app you can have code like:

#v+
# Module webui.py
import db

def get_users():
    cursor = db.getcon().cursor()
    cursor.execute("SELECT ...")
    return cursor.fetchall()

def format_users():
    buf = StringIO.StringIO()
    for row in get_users():
        print >>buf, ...
    return str(buf)
#v-

> One example that caught my attention is the round-up project but I
> have not yet managed decipher the code.

Another prominent example is PyPI, from the same author.

Gerhard
-- 
mail:   gh at ghaering.de
web:    http://ghaering.de/