I refer to this article about unit test of Hive SQL Queries : http://ift.tt/2ur8aEU
I found it very interesting regarding the fact that Big Data Use Case might always be tested, like with a "classic" Java Code.
Until this time, I was designing big queries from a client (DBViz for example), and when it was OK, I was preparing the query to be include in a OOZIE bundle (variabilization, ...) and most of the time, I forgot a point (a well named variable in the job.properties for example).
Well, what I want now is the ability to test my queries in the context of my OOZIE bundle before sending the code on the server and working by attempt...
So, I come to my question (at last ! :-)) : I'll be glad you give me your feedback about it and the way you do to be a good HiveQL developper ;)
Thanks.
Aucun commentaire:
Enregistrer un commentaire