def open(url=None, task=None):
from mongo_util import get_collection
query = son.SON(json.loads(url, object_hook=json_util.object_hook))
uri = query['inputURI']
uri_info = uri_parser.parse_uri(uri)
spec = query['query']
fields = query['fields']
skip = query['skip']
limit = query['limit']
timeout = query['timeout']
sort = query['sort']
slave_ok = query['slave_ok']
#go around: connect to the sonnection then choose db by ['dbname']
collection = get_collection(uri)
cursor = collection.find(spec=spec,
fields=fields,
skip=skip,
limit=limit,
sort=sort,
timeout=timeout,
slave_okay=slave_ok)
wrapper = MongoWrapper(cursor)
return wrapper
def __init__(self,stream,params):
from mongo_util import get_connection,get_collection
config = {}
for key, value in params.__dict__.iteritems():
config[key] = value
self.uri = config.get('output_uri')
self.conn = get_connection(self.uri)
self.coll = get_collection(self.uri)
self.key_name = config.get('job_output_key','_id')
self.value_name = config.get('job_output_value')
def test_traditional_way():
start = time.clock()
col = get_collection(config['input_uri'])
count = {}
cur = col.find()
for row in cur:
age = row['age']/10
if age in count:
count[age] += 1
else:
count[age] = 1
end = time.clock()
print "Time used: ", end-start
for key in count:
print key,count[key]
def open(url=None, task=None):
from mongo_util import get_collection
query = son.SON(json.loads(url, object_hook=json_util.object_hook))
uri = query['inputURI']
uri_info = uri_parser.parse_uri(uri)
spec = query['query']
fields = query['fields']
skip = query['skip']
limit = query['limit']
timeout = query['timeout']
sort = query['sort']
#go around: connect to the sonnection then choose db by ['dbname']
collection = get_collection(uri)
cursor = collection.find(spec = spec, fields = fields, skip = skip, limit = limit, sort = sort, timeout = timeout)
wrapper = MongoWrapper(cursor)
return wrapper
def calculate_unsharded_splits(config, slaveOk, uri):
"""@todo: Docstring for calculate_unsharded_splits
:returns: @todo
Note: collection_name seems unnecessary --CW
"""
splits = [] # will return this
logging.info("Calculating unsharded splits")
coll = get_collection(uri)
q = {} if not "query" in config else config.get("query")
# create the command to do the splits
# command to split should look like this VV
# SON([('splitVector', u'test.test_data'), ('maxChunkSize', 2),
# ('force', True), ('keyPattern', {'x': 1})])
split_key = config.get('split_key')
split_size = config.get('split_size')
full_name = coll.full_name
logging.info("Calculating unsharded splits on collection %s with Split Key %s" %
(full_name, split_key))
logging.info("Max split size :: %sMB" % split_size)
cmd = bson.son.SON()
cmd["splitVector"] = full_name
cmd["maxChunkSize"] = split_size
cmd["keyPattern"] = split_key
cmd["force"] = False
logging.debug("Issuing Command: %s" % cmd)
data = coll.database.command(cmd)
logging.debug("%r" % data)
# results should look like this
# {u'ok': 1.0, u'splitKeys': [{u'_id': ObjectId('4f49775348d9846c5e582b00')},
# {u'_id': ObjectId('4f49775548d9846c5e58553b')}]}
if data.get("err"):
raise Exception(data.get("err"))
elif data.get("ok") != 1.0:
raise Exception("Unable to calculate splits")
split_data = data.get('splitKeys')
if not split_data:
logging.warning("WARNING: No Input Splits were calculated by the split code. \
Proceeding with a *single* split. Data may be too small, try lowering \
'mongo.input.split_size' if this is undesirable.")
else:
logging.info("Calculated %s splits" % len(split_data))
last_key = None
for bound in split_data:
splits.append(_split(config, q, last_key, bound))
last_key = bound
splits.append(_split(config, q, last_key, None))
return [s.format_uri_with_query() for s in splits]
