Which of the following is the best way to migrate a large data load from one table to another, considering total time and blocking?

Question

Options

• ASplit the load size into many transactions.
• BSplit the load size in half and run simultaneously.
• CBatch into small loads and run in parallel.
• DBatch large loads into one transaction.

• A
  10% (4)
• B
  7% (3)
• C
  80% (33)
• D
  2% (1)

Explanation

Batching into small loads and running them in parallel (option C) is the optimal strategy because it minimizes lock duration per transaction (reducing blocking on the source/destination tables) while parallelism maximizes throughput and minimizes total elapsed time. Small transactions hold locks briefly, so other processes experience minimal contention. Option A (many transactions sequentially) reduces blocking per transaction but is slow because it runs serially. Option B (split in half, run simultaneously) reduces time but two large transactions still hold locks for extended periods, causing significant blocking. Option D (one large transaction) is the worst choice: a single massive transaction holds locks for the entire duration, blocking all other access to affected tables and risking long rollbacks if something fails.

No community discussion yet for this question.