Github user njayaram2 commented on a diff in the pull request:
https://github.com/apache/madlib/pull/315#discussion_r214485090
--- Diff: src/ports/postgres/modules/knn/knn.py_in ---
@@ -53,22 +55,12 @@ def knn_validate_src(schema_madlib, point_source,
point_column_name, point_id,
if label_column_name and label_column_name.strip():
cols_in_tbl_valid(point_source, [label_column_name], 'kNN')
- cols_in_tbl_valid(point_source, (point_column_name, point_id), 'kNN')
- cols_in_tbl_valid(test_source, (test_column_name, test_id), 'kNN')
-
- if not is_col_array(point_source, point_column_name):
- plpy.error("kNN Error: Feature column '{0}' in train table is not"
- " an array.".format(point_column_name))
- if not is_col_array(test_source, test_column_name):
- plpy.error("kNN Error: Feature column '{0}' in test table is not"
- " an array.".format(test_column_name))
--- End diff --
`point_column_name` and `test_column_name` params must be an array as this
if check suggests. If it's not an array it fails further down when the distance
function (such as `squared_dist_norm2`) is called.
I don't think the function `is_var_valid()` checks for these being arrays.
You may have to check them after the new asserts, using a new helper function
(`is_col_array()` cannot be used as is for expressions, and `is_var_valid()`
does not check for an array, but just the validity of the expression)
---
