def test_per_step_jobconf_can_set_number_of_tasks(self):
mr_job = MRTestPerStepJobConf(["-r", self.RUNNER, "--jobconf", "mapred.map.tasks=2"])
# need at least two items of input to get two map tasks
mr_job.sandbox(StringIO("foo\nbar\n"))
with mr_job.make_runner() as runner:
runner.run()
# sanity test: --jobconf should definitely work
self.assertEqual(runner.counters()[0]["count"]["mapper_init"], 2)
# the job sets its own mapred.map.tasks to 4 for the 2nd step
self.assertEqual(runner.counters()[1]["count"]["mapper_init"], 4)
def test_per_step_jobconf_can_set_number_of_tasks(self):
mr_job = MRTestPerStepJobConf([
'-r', self.RUNNER, '--jobconf', 'mapred.map.tasks=2',
])
# need at least two items of input to get two map tasks
mr_job.sandbox(BytesIO(b'foo\nbar\n'))
with mr_job.make_runner() as runner:
runner.run()
# sanity test: --jobconf should definitely work
self.assertEqual(runner.counters()[0]['count']['mapper_init'], 2)
# the job sets its own mapred.map.tasks to 4 for the 2nd step
self.assertEqual(runner.counters()[1]['count']['mapper_init'], 4)
def test_per_step_jobconf(self):
mr_job = MRTestPerStepJobConf([
'-r', self.RUNNER, '--jobconf', 'user.defined=something'])
mr_job.sandbox()
results = {}
with mr_job.make_runner() as runner:
runner.run()
for key, value in mr_job.parse_output(runner.cat_output()):
results[tuple(key)] = value
# user.defined gets re-defined in the second step
self.assertEqual(results[(0, 'user.defined')], 'something')
self.assertEqual(results[(1, 'user.defined')], 'nothing')
def test_per_step_jobconf(self):
mr_job = MRTestPerStepJobConf(["-r", self.RUNNER, "--jobconf", "user.defined=something"])
mr_job.sandbox()
results = {}
with mr_job.make_runner() as runner:
runner.run()
for line in runner.stream_output():
key, value = mr_job.parse_output_line(line)
results[tuple(key)] = value
# user.defined gets re-defined in the second step
self.assertEqual(results[(0, "user.defined")], "something")
self.assertEqual(results[(1, "user.defined")], "nothing")
def test_per_step_jobconf(self):
mr_job = MRTestPerStepJobConf([
'-r', self.RUNNER, '-D', 'user.defined=something'])
mr_job.sandbox()
results = {}
with mr_job.make_runner() as runner:
runner.run()
for key, value in mr_job.parse_output(runner.cat_output()):
results[tuple(key)] = value
# user.defined gets re-defined in the second step
self.assertEqual(results[(0, 'user.defined')], 'something')
self.assertEqual(results[(1, 'user.defined')], 'nothing')