Github user iyerr3 commented on a diff in the pull request:
https://github.com/apache/madlib/pull/237#discussion_r169846145
--- Diff: src/ports/postgres/modules/convex/mlp_igd.py_in ---
@@ -796,14 +807,34 @@ def mlp_predict(schema_madlib,
else:
# if not grouping, then directly read out the coeff, mean
# and std values from the model and standardization tables.
- standardization = plpy.execute(
- "SELECT * FROM {0}".format(standardization_table))[0]
+
+ # Fix to ensure that 1.12 models run on 1.13 or higher.
+ # As a result of adding grouping support in 1.13, the following
change
+ # was also made wrt standardization. The x_mean and x_std
+ # values were stored in the summary table itself in MADlib 1.12,
and
+ # they were named as: x_means and x_stds.
+ # From MADlib 1.13 onwards, these parameters were moved to the
+ # _standardization table, and were renamed to mean and std.
+ if is_v112_model:
--- End diff --
I feel we can simplify by creating two SQL strings and avoid all the
variables:
```
if pre_113:
SELECT x_means as mean, x_stds as std FROM {summary_table}
else:
SELECT mean, std FROM {standardization_table}
```
---
