ignore this

    #!/usr/bin/env python
    import sys
    import random
    from thrift import Thrift
    from thrift.transport import TSocket
    from thrift.transport import TTransport
    from thrift.protocol import TBinaryProtocol
    from hbase import Hbase
    from hbase.ttypes import *
    from datetime import datetime
    transport = TSocket.TSocket('localhost', 9090)
     
    # Buffering is critical. Raw sockets are very slow
    transport = TTransport.TBufferedTransport(transport)
     
    # Wrap in a protocol
    protocol = TBinaryProtocol.TBinaryProtocol(transport)
    client = Hbase.Client(protocol)
     
    transport.open()
     
    try: #Try to create the table if it doesn't exist
        client.createTable('test_table', [ColumnDescriptor(name='meta')])
    except AlreadyExists, tx:
        print "Thrift exception"
        print '%s' % (tx.message)
     
    for benchmark in range(8): #range(8) goes up to 10,000,000 rows, takes about 3hrs on my 2GB linode
        num = 10**benchmark
        print "Benchmark #%d: %d rows" %(benchmark, num)
        start = datetime.now()
        for i in range(num):
            client.deleteAllRow('test_table', "%d"%i)
        print '  Rows deleted: %s' % (datetime.now() - start)
        create_start = datetime.now()
        for i in range(num):
            client.mutateRow('test_table', "%d"%i,
                    [Mutation(column="meta:%s"%random.choice(['a', 'b', 'c']),
                        value="%d"%random.randrange(100))])
        print '  Rows created: %s' % (datetime.now() - create_start)
        get_start = datetime.now()
        r = client.getRow('test_table', "%s"%random.randrange(num))
        print '  Row lookup: %s' % (datetime.now() - get_start)
        scan_start = datetime.now()
        s = client.scannerOpen('test_table', '', ['meta:a'])
        r = client.scannerGet(s)
        count = 0
        r = client.scannerGet(s)
        while r:
            count += 1
            r = client.scannerGet(s)
        print '  Fetched %s rows with meta:A: %s' %(count, datetime.now() - scan_start)
        print '  Total Benchmark Time: %s' % (datetime.now() - start)
     
     
    # Sample output --
     
    #Benchmark #0: 1 rows
    #  Rows deleted: 0:00:00.000816
    #  Rows created: 0:00:00.001384
    #  Row lookup: 0:00:00.001078
    #  Fetched 319 rows with meta:A: 0:00:00.176335
    #  Total Benchmark Time: 0:00:00.179796
    #Benchmark #1: 10 rows
    #  Rows deleted: 0:00:00.004934
    #  Rows created: 0:00:00.006587
    #  Row lookup: 0:00:00.000725
    #  Fetched 316 rows with meta:A: 0:00:00.171574
    #  Total Benchmark Time: 0:00:00.183974
    #Benchmark #2: 100 rows
    #  Rows deleted: 0:00:00.051719
    #  Rows created: 0:00:00.064601
    #  Row lookup: 0:00:00.000687
    #  Fetched 328 rows with meta:A: 0:00:00.176993
    #  Total Benchmark Time: 0:00:00.294160
    #Benchmark #3: 1000 rows
    #  Rows deleted: 0:00:00.488865
    #  Rows created: 0:00:00.585270
    #  Row lookup: 0:00:00.000675
    #  Fetched 312 rows with meta:A: 0:00:00.168604
    #  Total Benchmark Time: 0:00:01.243603
    #Benchmark #4: 10000 rows
    #  Rows deleted: 0:00:04.546011
    #  Rows created: 0:00:05.570421
    #  Row lookup: 0:00:00.000995
    #  Fetched 3372 rows with meta:A: 0:00:01.810908
    #  Total Benchmark Time: 0:00:11.928531
    #Benchmark #5: 100000 rows
    #  Rows deleted: 0:00:41.258448
    #  Rows created: 0:00:55.093361
    #  Row lookup: 0:00:00.000818
    #  Fetched 33274 rows with meta:A: 0:00:17.005304
    #  Total Benchmark Time: 0:01:53.358131