Question

如果测试具有警告性,我就有一个马科探测器。警告涉及关系检查。如果存在警告,逻辑是另一种模式。

{% macro check_test_results() %}
  {% for result in results %}
    {% if result.status in ["warn", "fail"] %}
      {% if result.node.unique_id.startswith("test.prj_dbt.relationships_sch__fact_table_CUSTOMER_CODE__CUSTOMER_CODE__ref_sch___dim_customer_")
         or result.node.unique_id.startswith("test.prj_dbt.relationships_sch__fact_table_CUSTOMER_CODE__CUSTOMER_CODE__ref_sch___dim_product_") %}
        {{ log("Running exception query due to test failure: " ~ result.node.unique_id, info=True) }}
        
        -- Include your SQL directly here
        {% do run_query(
            " 
              select
                  f.*
              from mytable.FACT_SELLOUT f
              left join mytable.DIM_CUSTOMER c
              on f.CUSTOMER_CODE = c.CUSTOMER_CODE
              where c.CUSTOMER_CODE is null
            "
        ) %}
        
      {% endif %}
    {% endif %}
  {% endfor %}
{% endmacro %}

但是,如果出现这一错误,则KQ无法运行。此外,我更希望它执行另一个q子档案,而不是可能的话硬纸张。

 Encountered an error:
local variable  connection  referenced before assignment

Answer 1

The results variable available inside of the on-run-end hook context only (see details here). It s a significant dbt anti-pattern to run DDL operations (such as create table) outside of run, seed, snapshot operations. You should probably redesign your dbt models to achieve the desired behavior.

Most probably, you can do the the relationship check inside your model using SQL and joins and handle the output with CTE. It s hard to tell more without more details.

友情链接