python - GAE BigQuery 在开发服务器上运行，但部署时出现 HTTP 400 错误

Question

制作了一个 GAE 标准应用程序，将 .JSON 保存到 Google Cloud Storage，然后使用架构将 .JSON 加载到 BigQuery 中。通过 Google Cloud Shell 编写并运行它。

该脚本在开发服务器上启动时有效(保存的 .JSON，加载 BQ 表)。脚本在部署并访问 appspot URL 时出错。该错误是关于架构的......但它在开发上运行得很好。

部署并点击 appspot URL 时出错:

HttpError: https://www.googleapis.com/bigquery/v2/projects/api-gcs/jobs?alt=json returned "Required parameter: [resource.configuration.load.schema.field[0].name]">

代码:

import cloudstorage
from google.appengine.api import app_identity
from google.appengine.ext import blobstore
from google.appengine.ext.webapp import blobstore_handlers
import json
import sys
from collections import OrderedDict
import requests_toolbelt.adapters.appengine
requests_toolbelt.adapters.appengine.monkeypatch()
from oauth2client.client import GoogleCredentials
credentials=GoogleCredentials.get_application_default()
from googleapiclient.discovery import build
import os
import webapp2

# This handler creates a file in Cloud Storage using the cloudstorage
# client library and then reads the data back using the Blobstore API.
class CreateAndReadFileHandler(webapp2.RequestHandler):
    def get(self):
        dict_test = {'date': '2018-01-02', 'username': 'pasta456', 'age': 43, 'favorite_number': 1.22}

        bucket = app_identity.get_default_gcs_bucket_name()

        filename = '/{}/json_example.json'.format(bucket)

        with cloudstorage.open(filename, 'w') as filehandle:
            filehandle.write(json.dumps(dict_test))

        blobstore_filename = '/gs{}'.format(filename)
        blob_key = blobstore.create_gs_key(blobstore_filename)

        data = blobstore.fetch_data(blob_key, 0, 6)

        PROJECT = os.environ['PROJECT']
        BUCKET = os.environ['BUCKET']
        DATASET = os.environ['DATASET']

        service = build("bigquery", "v2", credentials = credentials)

        job = {
            "configuration": {
              "load": {
                "sourceUris": ["gs://XXX.appspot.com/json_example.json"],
                "schema": {
                    "fields" : [
                      {"name": "date",
                       "type": "DATE"},
                      {"name": "username",
                       "type": "STRING"},
                      {"name": "age",
                        "type": "INTEGER"},
                      {"name": "favorite_number",
                       "type": "FLOAT"}
                      ]
                },
                "destinationTable": {
                  "projectId": PROJECT,
                  "datasetId": DATASET,
                  "tableId": "json_test2"
                },
                "sourceFormat" : "NEWLINE_DELIMITED_JSON",
                "createDisposition": "CREATE_IF_NEEDED"
              }   
            }
          }


        response = service.jobs().insert(
                projectId = PROJECT,
                body = job
                ).execute()

Answer 1

答案:解决方案是架构的 JSON 格式。我仍然不知道为什么所讨论的模式格式可以在开发中工作，但在部署时却不能。下面是适用于部署的模式格式。

            "schema": {
                "fields" : [
                  {"name": "date2", "type": "DATE"},
                  {"name": "username", "type": "STRING"},
                  {"name": "age", "type": "INTEGER"},
                  {"name": "favorite_number", "type": "FLOAT"}
                  ]
            }