reductionista commented on a change in pull request #440: DL: Update training_preprocessor_dl to use bytea URL: https://github.com/apache/madlib/pull/440#discussion_r322484444
########## File path: src/ports/postgres/modules/deep_learning/test/input_data_preprocessor.sql_in ########## @@ -51,15 +51,12 @@ SELECT training_preprocessor_dl( SELECT assert(count(*)=4, 'Incorrect number of buffers in data_preprocessor_input_batch.') FROM data_preprocessor_input_batch; -SELECT assert(array_upper(independent_var, 2)=6, 'Incorrect buffer size.') +SELECT assert(independent_var_shape ='[1:4][1:6]', 'Incorrect buffer size.') FROM data_preprocessor_input_batch WHERE buffer_id=0; -SELECT assert(array_upper(independent_var, 1)=5, 'Incorrect buffer size.') +SELECT assert(independent_var_shape ='[1:5][1:6]', 'Incorrect buffer size.') Review comment: In test cases like these, it would be good to validate that the length of `independent_var` (in addition to checking that the `independent_var_shape` field is accurate). I think we could easily have a regression where the first ends up wrong but the second is right. For example, for these two cases we could do this by asserting `length(independent_var = 24` and `length(independent_var) = 30`. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
