Thanks, Shawn. That was my suspicion but I'm glad to hear it from someone else. Do you have any links or documentation I can give to a manager to help them feel more confident in this?
Also, I misspoke, I'm running Hive 1.2.1. Get Outlook for Android<https://aka.ms/ghei36> ________________________________ From: Shawn Weeks <swe...@weeksconsulting.us> Sent: Wednesday, November 6, 2019 5:35:17 PM To: user@hive.apache.org <user@hive.apache.org> Subject: Re: INSERT OVERWRITE Failure Saftey I’m not sure specific to Hive 1.3 but in other versions the data is written to a temp location and then at the end of the query the previous data is deleted and the new data is renamed/moved. Something to watch out for is if the query returns no rows than the old data isn’t removed. Thanks Shawn From: David M <mcginni...@outlook.com> Reply-To: "user@hive.apache.org" <user@hive.apache.org> Date: Wednesday, November 6, 2019 at 3:27 PM To: "user@hive.apache.org" <user@hive.apache.org> Subject: INSERT OVERWRITE Failure Saftey All, I have a Hive 1.3 cluster running in production, and there was a question about INSERT OVERWRITE queries on tables. If I perform an INSERT OVERWRITE query on a table, and the query fails half way through, will the old data still exist in the table? I’m not completely clear on the exact process which INSERT OVERWRITE follows, but I believe it puts the data into the staging folder, and then does a remove and move, which should be safe. It could also just wipe the folder before the query starts however, which would cause issues if the query itself failed. Can someone give me a definitive answer on this? Pointers to the source code or documentation that explains this would be even better. Thanks! David McGinnis
