I`m using AWS DMS to migrate data from a MySQL RDS to S3 with parquet format. After that I am running a GLUE crawler, creating a database and a table. Everything is working fine, but the timestamps.
As example , on AWS Athena query I'm getting the result: +51218-09-22 03:06:40.000 while the actual value is 2019-04-01-14.27.04 -0300. As you can see, the date is completely wrong. All other types were migrated fine
Has anyone had this problem?
Solution found here: https://forums.aws.amazon.com/thread.jspa?threadID=302156
Apparently, there is a mismatch between default timestamp that DMS migrate the data and how Athena map this timestamp (miliseconds/ nanoseconds).
"add parquetTimestampInMillisecond=true to the extra connection attributes of the target endpoint fixed this" then run DMS and Crawler again