Jelte Fennema-Nio
Jelte Fennema-Nio
Basically the same problem as #215, but for exensions instead of secrets. This was reported by @rakesh-tmdc in #209 I think we should probably do the same solution as in...
We completely broke COPY support without us noticing. That problem is being fixed by #264, but it might happen again if we don't add some tests. For the 0.1.0 release...
Right now only a Postgres superuser can use the extension. Any other user will get the following error: ``` ERROR: XX000: (PGDuckDB/CreatePlan) Executor Error: permission denied for sequence secrets_table_seq LOCATION:...
- [ ] how many threads - [ ] how much memory - [ ] disabled_filesystems = 'LocalFilesystem' (fixes #105) - [ ] more? Related to but separate from #216
DuckDB has a bunch of settings that users probably want to tune based on their specific usage. At least threads memory usage should be configurable, and probably more. I think...
Generally it's recommended to use the `timestamptz` type instead of the `timestamp` type, so we should support this. I think supporting this is probably very simple as it's stored the...
It's really easy to make critical security mistakes in the SQL scripts of an extension. Luckily [pgspot](https://github.com/timescale/pgspot) can help finding those.
After reverting the index scan support for #183 for 0.1.0 we'll want to add it back in a redesigned way for 0.2.0.
Running pg_duckdb in postgres instances that use a different encoding than UTF8 is bound to cause lots of problems. We should fail hard if we detect that this is the...
We should have some documentation on the official duckdb documentation site: https://duckdb.org/docs/