Environment details

• OS:
• Node.js version:
• npm version:
• @google-cloud/bigquery version: latest

Steps to reproduce

1. Pass startIndex to Job#getQueryResultsStream
2. First request to populate the stream is successful and starts at the correct index.
3. Second request fails with When using a page token, you cannot specify an arbitrary startIndex

It seems this is simply because startIndex should be used on the first fetch, and then discarded once page tokens can be used for the next query.