Database Connections in Python

A Python application connects to the database (creates a database handle) by calling the Database method connect(). The database must have been previously opened, possibly in a different process if the database is in shared memory. For example:

     
    db = exdb.open_database("myopendb", dict)
    con = db.connect();
     

The db.connect() returns a database connection that is used in subsequent calls to database APIs. Note that all applications must create a separate connection for each thread (task) that accesses the database; database connections cannot be shared between different threads or tasks.

Applications should disconnect from the database once the database connection is no longer needed to allow the database runtime to “de-allocate” internal memory used to maintain the connection. A Python application disconnects from the database by calling con.close() The database cannot be destroyed (closed) until all active connections are closed.

Example

Typical application code to connect to an in-memory database could look like the following:

     
    db = exdb.open_database("myopendb", dict)
    con = db.connect();
    ...
    con.close()
    db.close()
     

Asynchronous Distributed Connections

It is possible to process multiple SQL queries on database shards in parallel using the exdb.connect() method to connect to an AsyncDistributedSqlEngine. Then the execute() method is used to execute SQL statements on multiple shards. For example, the following code snippet creates a connection to an AsyncDistributedSqlEngine, creates SQL statement strings and arguments vectors, then executes the statements in parallel on 4 database shards and prints out the results.

 
    nWorkers = 4
     
    nodes = ("localhost:5001", "localhost:5002")
     
    conn = self.exdb.connect(nodes=nodes, async=True, nWorkers=nWorkers)
     
    q = ["INSERT INTO MyTable(pk, value, u2, i1, i2, i4, i8, flt, dbl) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)"]*nWorkers
     
    args = []
    for i in xrange(nWorkers):
        args.append((i*100, 'str%s' % i, i, i, i, i, i, i*100, i*100.9))
 
    r = conn.execute(q, args)
    print "len(r)=", len(r)
    for k in range(len(r)):
        pprint.pprint(r[k])