如何在特定的“完成”状态下获得工作

Question

我正在尝试确定各种工作状态。 Bigquery提供了我所知道的三种状态：DONE, PENDING, and RUNNING。但是，我正在尝试根据以下内容获取状态：

• 完成
• 有待
• 成功
• 错误
• 取消
• 运行

我怎么能以一种不太“昂贵”的方式做到这一点，因为我以一种“长轮询”的方式迭代大约100个结果，大约每十秒一次。目前我做的事情如下：

jobs = [job for job in self.bq_client.list_jobs(project=PROJECT_ID]
if state is not None:
    jobs = [job for job in jobs if job.state == state]

如果状态为DONE，RUNNING或PENDING，则上述情况有效。但我如何覆盖其他州？

Answer 1

州跟踪工作进度，如果您需要成功/失败信息，您想要查看errorResult的response。对于成功的工作，这将是None，对于取消的，你会得到{u'reason': u'stopped', u'message': u'Job execution was cancelled: User requested cancellation'}。我用来测试这个的代码：

from google.cloud import bigquery
client = bigquery.Client()

project = "[PROJECT-ID]"
states = ["RUNNING", "PENDING", "SUCCESSFUL", "CANCELLED", "FAILED"]


def returnState(job):
  if job.state == "DONE":
    if job.error_result is None:
      return "SUCCESSFUL"
    elif job.error_result['reason'] == u'stopped':
      return "CANCELLED"
    else:
      return "FAILED"
  else:
    return job.state


jobs = [job for job in client.list_jobs(project=project, max_results=10)]

for state in states:
  matching_jobs = [job for job in jobs if returnState(job) == state]

  for job in matching_jobs:
    print "Job ID: {0}, State: {1}, Error Result: {2}".format(job.job_id, state, job.error_result)

这将打印出如下内容：

$ python bq-status.py
Job ID: bquijob_..., State: SUCCESSFUL, Error Result: None
Job ID: bquijob_..., State: SUCCESSFUL, Error Result: None
Job ID: job_..., State: SUCCESSFUL, Error Result: None
Job ID: job_..., State: SUCCESSFUL, Error Result: None
Job ID: job_..., State: SUCCESSFUL, Error Result: None
Job ID: job_..., State: SUCCESSFUL, Error Result: None
Job ID: scheduled_query_..., State: SUCCESSFUL, Error Result: None
Job ID: bquijob_..., State: SUCCESSFUL, Error Result: None
Job ID: bquijob_..., State: CANCELLED, Error Result: {u'reason': u'stopped', u'message': u'Job execution was cancelled: User requested cancellation'}
Job ID: bquijob_..., State: FAILED, Error Result: {u'reason': u'invalidQuery', u'message': u'Syntax error: Illegal input character "\\\\" at [2:18]', u'location': u'query'}

请记住，加载作业可能会成功，但允许一些maxBadRecords，以便errorResult不会为空，等等。