You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the DuckDB JDBC driver's "jdbc:duckdb:/path/to/file" URL format only accepts files in the DuckDB file format. It would be useful if Parquet files would also be accepted as part of the connection URL.
Currently an exception is thrown: 'The file "/foo/bar/somefile.parquet" exists, but it is not a valid DuckDB database file!'
When connecting directly to a Parquet file with DuckDB's JDBC driver, DatabaseMetaData.getTables and related methods should probably act as if a single database table is available.
The use case is for external graphical tools (e.g. Ultorg in my case) that know how to work with arbitrary JDBC drivers, but which may not know about DuckDB's special "SELECT * FROM read_parquet('input.parquet');" syntax.
The text was updated successfully, but these errors were encountered:
That works! Direct support for Parquet files in the JDBC URL would still be a "nice to have", since it's friendlier to users who are not as familiar with command-line tools.
(In my case I'm actually developing the graphical tool in question, rather than being an end-user of the tool. Eventually I will probably make a custom connection UI that doesn't require users to deal with connection URLs at all... then perhaps I could automate the CREATE VIEW step, too.)
Currently, the DuckDB JDBC driver's "jdbc:duckdb:/path/to/file" URL format only accepts files in the DuckDB file format. It would be useful if Parquet files would also be accepted as part of the connection URL.
Currently an exception is thrown: 'The file "/foo/bar/somefile.parquet" exists, but it is not a valid DuckDB database file!'
When connecting directly to a Parquet file with DuckDB's JDBC driver, DatabaseMetaData.getTables and related methods should probably act as if a single database table is available.
The use case is for external graphical tools (e.g. Ultorg in my case) that know how to work with arbitrary JDBC drivers, but which may not know about DuckDB's special "SELECT * FROM read_parquet('input.parquet');" syntax.
The text was updated successfully, but these errors were encountered: