如何模拟对 pyspark sql 函数的内部调用 [英] How to mock inner call to pyspark sql function

问题描述

得到以下pyspark代码:

Got the following piece of pyspark code:

import pyspark.sql.functions as F

null_or_unknown_count = df.sample(0.01).filter(
    F.col('env').isNull() | (F.col('env') == 'Unknown')
).count()

在测试代码中，数据框被模拟，所以我试图为这个调用设置 return_value，如下所示:

In test code, the data frame is mocked, so I am trying to set the return_value for this call like this:

from unittest import mock
from unittest.mock import ANY

@mock.patch('pyspark.sql.DataFrame', spec=pyspark.sql.DataFrame)
def test_null_or_unknown_validation(self, mock_df):
    mock_df.sample(0.01).filter(ANY).count.return_value = 250

但这失败了:

File "/usr/local/lib/python3.7/site-packages/pyspark/sql/functions.py", line 44, in _
  jc = getattr(sc._jvm.functions, name)(col._jc if isinstance(col, Column) else col)
AttributeError: 'NoneType' object has no attribute '_jvm'

还尝试了 mock_df.sample().filter().count.return_value = 250，它给出了同样的错误.

Also tried mock_df.sample().filter().count.return_value = 250, which gives the same error.

我如何模拟过滤器，即 F.col('env').isNull() |(F.col('env') == 'Unknown') 正确吗?

How do I mock the filter i.e. F.col('env').isNull() | (F.col('env') == 'Unknown') correctly?

推荐答案

感谢我在工作中聪明的同事，这就是答案.我们必须模拟 pyspark.sql.functions.col 然后设置一个 return_value.

Thanks to my smart colleague at work, here is the answer. We have to mock pyspark.sql.functions.col and then set a return_value.

@mock.patch('pyspark.sql.functions.col')
@mock.patch('pyspark.sql.DataFrame', spec=pyspark.sql.DataFrame)
def test_null_or_unknown_validation(self, mock_df, mock_functions):
    mock_functions.isNull.return_value = True # (or False also works)
    mock_df.sample(0.01).filter(ANY).count.return_value = 250

使用 mock_df.sample().filter().count.return_value = 250 也可以正常工作.

Using mock_df.sample().filter().count.return_value = 250 also works fine.

这篇关于如何模拟对 pyspark sql 函数的内部调用的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！