Raise exception if user includes data product not in registry in config #46
Conversation
|
@DennisReddyhoff I haven't looked at the code, but it's okay not to raise an exception if a write doesn't exist - you may be writing a genuinely new thing. The corresponding working config should just give a version of 0.0.1 for PATCH, 0.1.0 for MINOR, etc. |
Oh yeah of course, in which case I'm not quite sure I understand the issue @soniamitchell raised |
|
It may be a typo. You should raise an exception if a version does not exist in a read block, and you should raise an exception if you specify an exact version and it does exist in a write block. |
|
I'm afraid I can't remember exactly what happened, I'll provide more detail next time. I've just tested a data product that doesn't exist in a |
|
I think that's covered in #44? This was about when |
|
Should it not give an error if the data product doesn’t exist? Otherwise how will the user know? |
|
Oh, sorry - yes. It was the comment about the write generating errors that I was concerned about that you mentioned in #34 - it's okay for a data product to not exist there, but you should raise an exception if you specify an exact version and it does exist in a write block. But maybe I misunderstood the issue you were raising. |
|
Ok, for |
|
on it |
|
perfect! |
|
Actually, one tiny comment. When reading a data product (that does exist) of a version that doesn't exist, the error message could list the version number like it does in the |
|
Ok cool, will fix that and merge |
|
Kudos, SonarCloud Quality Gate passed!
|
An exception should now be raised if user includes a
data_productinwriteorrunblocks that doesn't exist in the registry. Issue #34 @soniamitchell