3 ms·
OP's frustrations are quite understandable but I can think of a couple of areas where ORM-like concepts have worked well. PyDAL[0] and Linq are both incredibly
by tomcam 8y ago
OP's frustrations are quite understandable but I can think of a couple of areas where ORM-like concepts have worked well. PyDAL[0] and Linq are both incredibly expressive and PyDAL is also low overhead (it's not an ORM, it's a data access language, which means it doesn't have the baggage of an object-oriented layer).
Another reason to use a DAL like PyDAL is that, properly written, it can abstract away the data layer such that you can switch from (for example) MySQL to PostGres to Google Cloud SQL with essentially no retooling required.
Finally, using a DAL can help keep inexperienced programmers on track with security, accessibility, and performance requirements.
[0] https://github.com/web2py/pydal https://github.com/web2py/pydal
- tannhaeuser 8y ago> Finally, using a DAL can help keep inexperienced programmers on track with security, accessibility, and performance requirements. This is a sure way they'll never learn the real thing. I know since I'm sitting next to a (young but talented) guy who knows only Spring and JPA, and freaks out all the time because the developer of the legacy system had dared to assemble SQL queries manually in the code (which are complex enough I personally hadn't bothered using JPQL in a trial-and-error development model either). Wasn't the post exactly about not wasting novice and other developer's time with crap criteria APIs and such? Just compare the sheer size of the SQL spec or reference manual against the shallow description of JPQL to judge which language you should be using for a project with any kind of depth.