Here is a table example that would have caused the issue:

 

CREATE TABLE ivan.a (a int NOT NULL, b int NOT NULL, c varchar(20) NOT NULL, d date NOT NULL, PRIMARY KEY(a))

ORDER BY a

UNSEGMENTED ALL NODES

PARTITION BY a.b

GO

SELECT distinct t.table_schema, t.table_name, t.is_temp_table, t.partition_expression, c.column_name, c.data_type, c.is_nullable, c.column_default, c.is_identity, p.node_name, p.is_segmented, pc.sort_position, pc.encoding_type, pc.access_rank, s.session_cache_count, s.increment_by, s.minimum, pc.column_position=pc.sort_position, c.ordinal_position, false FROM v_catalog.tables t LEFT JOIN v_catalog.columns c ON t.table_id = c.table_id LEFT JOIN v_catalog.projections p ON (t.table_id = p.anchor_table_id AND t.table_name = p.projection_basename) LEFT JOIN v_catalog.projection_columns pc ON (p.projection_name = pc.projection_name AND c.column_name = pc.projection_column_name AND pc.table_schema = t.table_schema AND pc.table_name = t.table_name) LEFT JOIN v_catalog.sequences s ON t.table_id = s.identity_table_id and c.is_identity = true WHERE t.table_name = 'a' AND t.table_schema = 'ivan' ORDER BY c.ordinal_position 

GO

 

We now handle duplicate columns on client side and unsegmented info on client side.  Scripting should work correctly now.