monetdb full outer join resulting in varchar type_digits=0 - monetdb

Question

I am using MonetDB v11.29.7 "Mar2018-SP1" on a Windows10 x64 bit operating system. When I perform a full outer join with two tables on respective varchar columns with lengths > 0 (type_digits > 0), the resultant column in the target table yields a varchar column with type_digits=0, although the column data seems to display the proper, non-null varchar records.
I am not sure how to interpret column information of type=varchar and type_digits=0. This state is causing issues in the subsequent handling/extraction of data via Python interfaces (UDFs), as the expected Python dtype for the data of this column is ambiguous for Python numpy conversion.
I have provided a simple example whereby I created two small tables (dummy4 and dummy5) with two columns each and then create a third table (dummy6) using a full outer join command.
For table dummy6 and column "key", I would have expected the type_digits=32 (as per the "key" columns in the two source tables dummy4 & dummy5). Additionally, how should I interpret type=varchar and type_digits=0 state? What would be the proper handling/expectation when accessing/allocating a Python/numpy array for extracting the "key" column of table "dummy6" (via Python UDFs) in this case?
create table dummy4(key varchar(32), val int);
insert into dummy4 values('AAAAAAAA',1);
insert into dummy4 values('BBBBBBBBB',2);
select * from dummy4;
+-----------+------+
| key | val |
+===========+======+
| AAAAAAAA | 1 |
| BBBBBBBBB | 2 |
+-----------+------+
create table dummy5(key varchar(32), val int);
insert into dummy5 values('CCCCCCCC',3);
insert into dummy5 values('DDDDDDDD',4);
select * from dummy5;
+----------+------+
| key | val |
+==========+======+
| CCCCCCCC | 3 |
| DDDDDDDD | 4 |
+----------+------+
create table dummy6 as select key, dummy4.val as "val4", dummy5.val as "val5" from dummy4 full outer join dummy5 using (key);
select * from dummy6;
+-----------+------+------+
| key | val4 | val5 |
+===========+======+======+
| AAAAAAAA | 1 | null |
| BBBBBBBBB | 2 | null |
| CCCCCCCC | null | 3 |
| DDDDDDDD | null | 4 |
+-----------+------+------+
select t.name as "table_name", t.id as "table_id", c.id as "column_id", c.name as "column_name", c.type, c.type_digits from sys.tables t JOIN sys.columns c ON c.table_id = t.id where t.name = 'dummy4';
+------------+----------+-----------+-------------+---------+-------------+
| table_name | table_id | column_id | column_name | type | type_digits |
+============+==========+===========+=============+=========+=============+
| dummy4 | 78445 | 78443 | key | varchar | 32 |
| dummy4 | 78445 | 78444 | val | int | 32 |
+------------+----------+-----------+-------------+---------+-------------+
select t.name as "table_name", t.id as "table_id", c.id as "column_id", c.name as "column_name", c.type, c.type_digits from sys.tables t JOIN sys.columns c ON c.table_id = t.id where t.name = 'dummy5';
+------------+----------+-----------+-------------+---------+-------------+
| table_name | table_id | column_id | column_name | type | type_digits |
+============+==========+===========+=============+=========+=============+
| dummy5 | 78449 | 78447 | key | varchar | 32 |
| dummy5 | 78449 | 78448 | val | int | 32 |
+------------+----------+-----------+-------------+---------+-------------+
select t.name as "table_name", t.id as "table_id", c.id as "column_id", c.name as "column_name", c.type, c.type_digits from sys.tables t JOIN sys.columns c ON c.table_id = t.id where t.name = 'dummy6';
+------------+----------+-----------+-------------+---------+-------------+
| table_name | table_id | column_id | column_name | type | type_digits |
+============+==========+===========+=============+=========+=============+
| dummy6 | 78457 | 78454 | key | varchar | 0 |
| dummy6 | 78457 | 78455 | val4 | int | 32 |
| dummy6 | 78457 | 78456 | val5 | int | 32 |
+------------+----------+-----------+-------------+---------+-------------+

Answer 1

In fact this was a MonetDB's bug and was fixed today. Th fix will be featured on the upcoming Nov2019 release.