BigQuery: need to return unique values for Record Type in a correlated subquery

问题

I am having a big query table in this format

I have the record values for the table in this format :

The same Instance Id and project Id may be repeated any no of times with different time series. I am trying for a select statment which will return one value(no cartestian product) for each row

output like

 InstanceId    ProjectId Time
    
    2763333     manage-x 10:30
    2763333     manage-x 11:30
    2763334     manage-y 10:30

since this is a record type, I tried this the table name is metric

select res.value from metric,unnest(resource.labels) as res where res.key="instance_id"

This gives me the corect value of 2763339646023081

Now I want to get project_id in the same statement so I need a correlated subquery for that similar to sql

select res.value from metric,unnest(resource.labels) as res,(select proj.value from metric,unnest(resource.labels) as proj where proj.key="project_id" and this part i need help to refer the res.value(instance_id) from the outer query to match to the corresponding instance for the project_id in the inner query)) where res.key="instance_id"

So as shown above I am not sure how to refer the instance_id in the inner subquery to get the corresponding project_id, I am very new to BigQuery and I have tried various combinations of joins yet it does not work and gives me cartestian products. Please give your suggestions and help. Thanks

回答1:

Below is for BigQuery Standard SQL

#standardSQL
SELECT 
  (SELECT value FROM UNNEST(resource.labels) AS res WHERE key = 'instance_id') AS instance_id,
  (SELECT value FROM UNNEST(resource.labels) AS res WHERE key = 'project_id') AS project_id
FROM `project.dataset.metric`

来源：https://stackoverflow.com/questions/62913490/bigquery-need-to-return-unique-values-for-record-type-in-a-correlated-subquery

标签

sql

google-bigquery