Environment details

• OS: macOS 13.4
• Node.js version: 16.20
• npm version: 8.19.4
• @google-cloud/bigquery version: 6.2.0

Steps to reproduce

1. Run the following script:

   import { BigQuery } from '@google-cloud/bigquery';
   
   const credentials = {
     /* service_credentials for a given project */
   };
   
   const bigquery = new BigQuery({
     credentials,
     projectId: credentials.project_id,
   });
   
   (async () => {
     const [job] = await bigquery.createQueryJob({
       query: 'SELECT 1 AS col',
     });
     const [rows, , response] = await job.getQueryResults({ maxResults: 1 });
     console.log(rows);
     console.log(response);
     const [rows2, , response2] = await job.getQueryResults();
     console.log(rows2);
     console.log(response2);
   })()
     .then(() => console.log('done'))
     .catch((err) => console.error(err));

2. See that response prints out an object that can be used, while response2 is undefined. Order to the calls does not matter

While both response and response2 are typed as bigquery.IGetQueryResultsResponse | undefined, it's non-obvious from reading the docs as to when it should be defined vs not.

My use case was wanting to reference response2.schema and finding it not defined, and so having to run this initial fetch first to get the schema metadata, and then a second request to get all the rows.