What happens?

Persistent .duckdb database exist, in which two views are defined:

• first_view is defined as a select over a read_parquet(...) function
• second_view is defined as a view on top of first_view

Both these views can be queried when the .duckdb database is opened with the CLI tool.

After using ATTACH from another database (eg :memory) attempts to query second_view fails with:

CatalogException
This cell raised an exception: CatalogException('Catalog Error: Table with name first_view does not exist! Did you mean "demo.first_view"?')

Details:

Traceback (most recent call last):
  File "C:\Users\jbi\Repos\marimo\bugreport\.venv\Lib\site-packages\marimo\_runtime\executor.py", line 170, in execute_cell
    exec(cell.body, glbls)
  Cell marimo://C:\Users\jbi\Repos\marimo\bugreport\notebooks\Myname.py#cell=cell-3, line 1, in <module>
    _df = mo.sql(
          ^^^^^^^
  File "C:\Users\jbi\Repos\marimo\bugreport\.venv\Lib\site-packages\marimo\_sql\sql.py", line 40, in sql
    relation = duckdb.sql(query=query)
               ^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\jbi\Repos\marimo\bugreport\.venv\Lib\site-packages\duckdb\__init__.py", line 457, in sql
    return conn.sql(query, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
duckdb.duckdb.CatalogException: Catalog Error: Table with name first_view does not exist!
Did you mean "demo.first_view"?

To Reproduce

DuckDB CLI is first used to create a .parquet file. Start CLI in memory mode:

duckdb

Execute statements:

CREATE TABLE mytable (id INTEGER PRIMARY KEY, j VARCHAR);
INSERT INTO mytable VALUES (1, 'a');
INSERT INTO mytable VALUES (b, '2');

COPY
  (SELECT * FROM mytable)
  TO 'demo.parquet'
  (FORMAT 'parquet');

Exit CLI tool.

Next step is to create the persistent .duckdb database. Open/create DB:

duckdb demo.duckdb

Execute statements to create view. Note that full-path to demo.parquet file is required.

CREATE VIEW first_view AS SELECT id, j FROM read_parquet('C:\Temp\demo.parquet');
CREATE VIEW second_view AS SELECT * FROM first_view;

Edit by @szarnyasg – Bash one-liner:

duckdb demo.duckdb -c "CREATE VIEW first_view AS SELECT id, j FROM read_parquet('`pwd`/demo.parquet'); CREATE VIEW second_view AS SELECT * FROM first_view;"

These views can both be queried as one would expect.

Exit CLI tool, and open new a transient duck-db database:

duckdb

Attach the .duckdb database

ATTACH 'demo.duckdb';

Query of first_view works:

SELECT * FROM demo.main.first_view;

Query of second_view fails:

SELECT * FROM demo.main.second_view;

Changing the namespace makes the second query work:

USE demo.main;

OS:

Windows 11 x64

DuckDB Version:

1.0.0

DuckDB Client:

duckdb cli

Full Name:

Jannick Bitsch

Affiliation:

Danske Commodities A/S

What is the latest build you tested with? If possible, we recommend testing with the latest nightly build.

I have tested with a stable release

Did you include all relevant data sets for reproducing the issue?

Yes

Did you include all code required to reproduce the issue?

• Yes, I have

Did you include all relevant configuration (e.g., CPU architecture, Python version, Linux distribution) to reproduce the issue?

• Yes, I have