Azure 函数不执行其他 Python 函数

Question

我在main.py 中有以下代码，这是我项目的简化版本，我在其中检索一些数据、清理数据并将其导出到 Azure 中的 SQL 数据库。

问题是，当我使用 Azure 函数在本地运行它时，一切正常，我可以在终端中看到日志记录。

但是当我用azure start --verbose 运行它时，函数可以正确启动，但是执行我的函数export_data 不起作用。

简化代码：

# import packages
# load_dotenv()

def get_data():
    # function which retrieves data


def export_data():
    # function which exports data and uses environment variables


def main(timer: func.TimerRequest) -> None:
    utc_timestamp = datetime.datetime.utcnow().replace(
        tzinfo=datetime.timezone.utc).isoformat()

    if timer.past_due:
        logging.info('The timer is past due! Writing data to database')
        export_data()

    logging.info('Python timer trigger function ran at %s', utc_timestamp)


if __name__ == "__main__":
    export_data()

我的代码：

import os
import datetime
import logging
import pandas as pd
from sqlalchemy import create_engine
from dotenv import load_dotenv
import azure.functions as func
load_dotenv()


logging.basicConfig(
    format="%(asctime)s.%(msecs)03d [%(levelname)-5s] [%(name)s] - %(message)s",
    datefmt="%Y-%m-%d %H:%M:%S",
    level=logging.INFO,
)


def get_data():
    # get data
    logging.info("retrieving data")
    url = "https://data.rivm.nl/covid-19/COVID-19_aantallen_gemeente_per_dag.csv"
    df = pd.read_csv(url, sep=";")

    return df


def export_data():
    # export data
    df = get_data()

    uid = os.environ.get("username")
    password = os.environ.get("pw")
    server = '***.database.windows.net'
    database = os.environ.get("database")
    driver = "ODBC Driver 17 for SQL Server"

    connection_string = (
        f"mssql+pyodbc://{uid}:{password}@" f"{server}:1433/{database}?driver={driver}"
    )

    engine = create_engine(connection_string)
    con = engine.connect()

    logging.info("exporting data")
    df.head(10).to_sql('covid_19', con=con, if_exists='replace')


def main(timer: func.TimerRequest) -> None:
    utc_timestamp = datetime.datetime.utcnow().replace(
        tzinfo=datetime.timezone.utc).isoformat()

    if timer.past_due:
        logging.info('The timer is past due! Writing data to database')
        export_data()

    logging.info('Python timer trigger function ran at %s', utc_timestamp)


if __name__ == "__main__":
    export_data()

function.json：

{
  "scriptFile": "main.py",
  "bindings": [
    {
      "name": "timer",
      "type": "timerTrigger",
      "direction": "in",
      "schedule": "0 */5 * * * *"
    }
  ]
}

---

编辑

所以这是日志的一部分，它显示了当我最初使用 func start --verbose 调用函数时函数运行的情况。注意时间，它不会在整个 5 分钟内发生，因为我的日程安排是基于此 ("schedule": "0 */5 * * * *")：

[2020-10-23T09:23:43.339] The timer is past due! Writing data to database
[2020-10-23T09:23:43.340] retrieving data
[2020-10-23T09:23:45.323] exporting data
[2020-10-23T09:23:46.182] Python timer trigger function ran at 2020-10-23T09:23:43.332697+00:00

那么在这部分之后，函数会根据定时器触发运行，但是什么也没有发生：

[2020-10-23T09:25:00.009] Executing 'Functions.zyppcovid' (Reason='Timer fired at 2020-10-23T11:25:00.0076610+02:00', Id=***)
[2020-10-23T09:25:00.012] Received FunctionInvocationRequest, request ID: ***, function ID: ***, invocation ID: ***
[2020-10-23T09:25:00.013] Function is sync, request ID: ***,function ID: ***, invocation ID: ***
[2020-10-23T09:25:00.013] Python timer trigger function ran at 2020-10-23T09:25:00.012563+00:00
[2020-10-23T09:25:00.013] Successfully processed FunctionInvocationRequest, request ID: ***, function ID: ***, invocation ID: ***
[2020-10-23T09:25:00.014] Executed 'Functions.zyppcovid' (Succeeded, Id=***, Duration=6ms)

Answer 1

代码已经有一个定时器触发的主函数def main(timer: func.TimerRequest) -> None:。所以它不能做另一个主要的if __name__ == "__main__":。我这边测试一下，如果我只是用main创建一个简单的python代码，它可以在main函数成功下做export_data()。

Answer 2

您应该将 export_data() 从 if 语句中取出。 if 语句只有在计时器过期时才会执行。

您很可能希望在每次运行 azure 函数时都执行 export_date()。

试试这个（注意用户定义的函数不再在 if 条件语句中，现在应该在函数运行时运行，而不仅仅是在计时器过期时）：

import os
import datetime
import logging
import pandas as pd
from sqlalchemy import create_engine
from dotenv import load_dotenv
import azure.functions as func
load_dotenv()


logging.basicConfig(
    format="%(asctime)s.%(msecs)03d [%(levelname)-5s] [%(name)s] - %(message)s",
    datefmt="%Y-%m-%d %H:%M:%S",
    level=logging.INFO,
)


def get_data():
    # get data
    logging.info("retrieving data")
    url = "https://data.rivm.nl/covid-19/COVID-19_aantallen_gemeente_per_dag.csv"
    df = pd.read_csv(url, sep=";")

    return df


def export_data():
    # export data
    df = get_data()

    uid = os.environ.get("username")
    password = os.environ.get("pw")
    server = '***.database.windows.net'
    database = os.environ.get("database")
    driver = "ODBC Driver 17 for SQL Server"

    connection_string = (
        f"mssql+pyodbc://{uid}:{password}@" f"{server}:1433/{database}?driver={driver}"
    )

    engine = create_engine(connection_string)
    con = engine.connect()

    logging.info("exporting data")
    df.head(10).to_sql('covid_19', con=con, if_exists='replace')


def main(timer: func.TimerRequest) -> None:
    utc_timestamp = datetime.datetime.utcnow().replace(
        tzinfo=datetime.timezone.utc).isoformat()
    
    logging.info('Writing data to database')
    export_data()
    logging.info('Writing data to database complete')

    if timer.past_due:
        logging.info('The timer is past due!')
        

    logging.info('Python timer trigger function ran at %s', utc_timestamp)