Your organization has been collecting and analyzing data in Google BigQuery for 6 months. The majority of the data analyzed is placed in a time-partitioned table named events_partitioned. To reduce…

Question

Options

• ACreate a new view over events using standard SQL
• BCreate a new partitioned table using a standard SQL query
• CCreate a new view over events_partitioned using standard SQL
• DCreate a service account for the ODBC connection to use for authentication
• ECreate a Google Cloud Identity and Access Management (Cloud IAM) role for the ODBC connection and shared "events"

• A
  17% (4)
• B
  4% (1)
• C
  74% (17)
• E
  4% (1)

Explanation

C and D address the two distinct problems introduced by ODBC: dialect incompatibility and authentication. The BigQuery ODBC driver only supports standard SQL - it cannot execute legacy SQL, so the existing events view would fail at query time. You must create a new view directly over events_partitioned (not over events) because a standard SQL view cannot reference a legacy SQL view. For authentication, ODBC connections are application-level, non-interactive clients, making a service account the correct credential mechanism - not a human user login.

A is wrong for the same reason as above: standard SQL cannot wrap a legacy SQL view, so building events_over_events would still break.
B is wrong because duplicating data into a new partitioned table is unnecessary and costly - a standard SQL view achieves the same result without data movement.
E is wrong because an IAM role defines permissions, not credentials; you'd assign an IAM role to the service account, but the role alone cannot authenticate a connection.

Memory tip: ODBC = Standard SQL + Service account - two S's. Whenever an application (non-human) needs to connect to BigQuery, always pair a new standard SQL object with a service account.

No community discussion yet for this question.