I am inserting batches in the RT index, but i have a rare situation, when my insert fails because of duplicate id in the index. If i replace my “INSERT INTO” to “REPLACE INTO” and will always use replace command, will i have a performance drop in query, if there is no duplicates in batch?
You can try it using Manticore Load Emulator, but I think the answer is no - if there are no duplicates, REPLACE should be just as fast as INSERT.
1 Like