My team is considering utilizing Hibernate/Envers for a granular point in time rollback solution for Redshift data. I have heard that using an ORM with Redshift is not a good idea. Are there particular drawbacks related to using an ORM such as hibernate with Redshift?

有帮助吗?

解决方案

There are known drawbacks related to using ORMs, period. They are a result of the inherent impedance between the object-oriented and relational models of the world. These issues will probably be exacerbated by the lack of Redshift-specific Hibernate dialect, and using a Postgres dialect will likely cause additional problems, as many Postgres features are not supported by Redshift.

In short, my opinionated advice would be, don't use an ORM for any non-trivial application, Redshift or not. "[G]ranular point in time rollback solution for Redshift data" doesn't sound like one.

许可以下: CC-BY-SA归因
不隶属于 dba.stackexchange
scroll top