kaknikhil commented on a change in pull request #361: Minibatch Preprocessor
DL: Add optional num_classes param.
URL: https://github.com/apache/madlib/pull/361#discussion_r271051530
##########
File path:
src/ports/postgres/modules/utilities/test/minibatch_preprocessing_dl.sql_in
##########
@@ -205,3 +210,81 @@ SELECT assert (dependent_vartype = 'integer[]' AND
class_values IS NULL,
'Summary Validation failed. Actual:' || __to_char(summary)
) from (select * from minibatch_preprocessor_dl_batch_summary)
summary;
+
+-- Test cases with NULL in class values
+DROP TABLE IF EXISTS minibatch_preprocessor_dl_input_null;
+CREATE TABLE minibatch_preprocessor_dl_input_null(id serial, x double
precision[], label TEXT);
+INSERT INTO minibatch_preprocessor_dl_input_null(x, label) VALUES
+(ARRAY[1,2,3,4,5,6], 'a'),
+(ARRAY[11,2,3,4,5,6], 'a'),
+(ARRAY[11,22,33,4,5,6], NULL),
+(ARRAY[11,22,33,44,5,6], 'a'),
+(ARRAY[11,22,33,44,65,6], 'a'),
+(ARRAY[11,22,33,44,65,56], 'a'),
+(ARRAY[11,22,33,44,65,56], 'a'),
+(ARRAY[11,22,33,44,65,56], NULL),
+(ARRAY[11,22,33,44,65,56], 'a'),
+(ARRAY[11,22,33,44,65,56], 'a'),
+(ARRAY[11,22,33,44,65,56], NULL),
+(ARRAY[11,22,33,44,65,56], 'a'),
+(ARRAY[11,22,33,144,65,56], 'b'),
+(ARRAY[11,22,233,44,65,56], 'b'),
+(ARRAY[11,22,33,44,65,56], 'b'),
+(ARRAY[11,22,33,44,65,56], 'b'),
+(ARRAY[11,22,33,44,65,56], NULL);
+
+DROP TABLE IF EXISTS minibatch_preprocessor_dl_batch,
minibatch_preprocessor_dl_batch_summary;
+SELECT minibatch_preprocessor_dl(
+ 'minibatch_preprocessor_dl_input_null',
+ 'minibatch_preprocessor_dl_batch',
+ 'label',
+ 'x',
+ 4,
+ 5,
+ 5 -- num_classes
+ );
+-- Test summary table if class_values has NULL as a legitimate
+-- class label, and also two other NULLs because num_classes=5
+-- but table has only 3 distinct class labels (including NULL)
+SELECT assert
Review comment:
+1 for adding this test
----------------------------------------------------------------
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
