Details
-
Bug
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
None
-
None
-
None
Description
While working on https://issues.apache.org/jira/browse/MADLIB-1128, we found a bug in the correlation/covariance module.
Repro steps:
create table foo(i int , j int); insert into foo values(1,NULL); madlib=# select madlib.correlation('foo','foo_out'); ERROR: plpy.SPIError: Function "madlib.correlation_transition(double precision[],double precision[],double precision[])": Correlation: Mean vector contains NULL. (UDF_impl.hpp:210) (entry db 127.0.0.1:15432 pid=46000) CONTEXT: Traceback (most recent call last): PL/Python function "correlation", line 23, in <module> return correlation.correlation(**globals()) PL/Python function "correlation", line 71, in correlation PL/Python function "correlation", line 207, in _populate_output_table PL/Python function "correlation" madlib=#
This was introduced in https://issues.apache.org/jira/browse/MADLIB-1166 when we started supporting null values.