DuckDBPyRelation.create_view cannot create a view in another schema or database

Offen
  • Stimmt etwas nicht?
#753 0 Kommentare 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen

Maintainer antworten meist innerhalb von 1 Tag

Dieses Issue hat noch niemand übernommen.

Bewertung

Schwierigkeit
3/5
Geschätzter Aufwand
1-2 Tage
Anfängerfreundlichkeit
54/100
Issue-Typ
Feature
Klarheit
Größtenteils klar
Aktivitätsstatus
Aktiv
Tech-Stack
python, sql
Bereich
api, database

Rechercherichtung

Start at the public DuckDBPyRelation.create_view and to_view implementations and trace how they pass the view name to the connection; the issue identifies private _create_view and _render as a working reference. Reproduce with the supplied script, including a qualified schema name and attached database, then check that the public API can create views at those paths while preserving relation parameters and referenced Python objects.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Beschreibung

We'd like to use rel.create_view("schema_name.view_name") to create the view view_name under the schema schema_name, but currently have to resort to calling the private methods and functions con._create_view and relation._render.

[!NOTE]
The following content was generated by AI.

What happens?

DuckDBPyRelation.create_view(view_name) (and to_view) takes one identifier. A qualified name is not split: rel.create_view("s.v") creates a view literally named "s.v" in memory.main, not a view v in schema s. There is no public way to create a view of a relation in another schema, another attached database, or a nested schema (DuckDB 2.0).

The workarounds lose what create_view gives. CREATE VIEW s.v AS {rel.sql_query()} reads the Python objects the relation scans by an internal name, so it fails once the relation is gone ("This query reads a pandas DataFrame or NumPy array that was released"), and it breaks on a relation with parameters ("Values were not provided for the following parameters"). The private con._create_view(path, sql) with _render(rel, literals=True, wrap_text=True) does exactly the right thing and already accepts a tuple path, nested schemas included.

We use that private call in datarecord to register a record's relations as views (<db>.<name>.dims.<dim>) that SQL clients read over a DuckDB server, and have to cap duckdb at the release we tested because of it.

Could create_view / to_view accept a qualified name, for example a tuple ("s", "nested", "v") like _create_view, or schema= / catalog= arguments?

To Reproduce
import duckdb
from duckdb._relation import _render

con = duckdb.connect()
con.sql("CREATE SCHEMA s; CREATE SCHEMA s.nested; ATTACH ':memory:' AS other")
rel = con.sql("SELECT 42 AS x")

rel.create_view("s.v")
rel.create_view("other.main.v")
print(con.sql(
    "SELECT database_name, schema_name, view_name FROM duckdb_views() "
    "WHERE NOT internal ORDER BY ALL"
).fetchall())
# [('memory', 'main', 'other.main.v'), ('memory', 'main', 's.v')]
# expected: ('memory', 's', 'v') and ('other', 'main', 'v')

con._create_view(("s", "nested", "w"), _render(rel, literals=True, wrap_text=True)[0])
print(con.sql("FROM s.nested.w").fetchall())  # [(42,)]: the private call does it
OS:

Linux x86_64

DuckDB Package Version:

2.0.0

Python Version:

3.14.6

Full Name:

Jonas Hörsch

Affiliation:

Open Energy Transition

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 (2.0.0)

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

Not applicable - the reproduction does not require a data set

Did you include all code required to reproduce the issue?
  • Yes, I have
Did you include all relevant configuration to reproduce the issue?
  • Yes, I have
Vorherrschende Sprache
Python
Sterne
203
Forks
124
Ø Merge
14 Std. 14 Min.
Gemergte PRs (30 T.)
70

Entwicklungsumgebung

Erste Schritte

  1. Lesen Sie das ganze Issue und danach den Beitragsleitfaden des Projekts.
  2. Schreiben Sie ins Issue, dass Sie es übernehmen — das erspart doppelte Arbeit.
  3. Forken Sie das Repository und arbeiten Sie in einem Branch.
  4. Öffnen Sie einen Pull Request, der die Issue-Nummer nennt.

Mehr aus duckdb/duckdb-python

Alle Issues in duckdb/duckdb-python

Ähnliche Issues

Weitere Issues zu Python

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.