Thijs

> I'm not sure if this is the same issue, but this also produces the error message: > > ```sql > create table a (i int primary key, j int);...

Fixed by https://github.com/duckdb/duckdb/pull/12226

I wouldnt say the fix has been merged The exception/crash is resolved, the functionality is still missing

To get the ball rolling, this is caused by the metadata block being located at the end. Current limitation: This is never moved after it is first created, so even...

Probably mentioning `pragma_metadata_info` is handy as it can show the block id of the metadata block, which can confirm that the user is suffering from this limitation Using COPY FROM...

> My preferred way out would be probably to have an option to specify a list of extensions to be AutoLoaded during Initialization. This is enabled through `EXTENSION_SETTINGS` in `extension_entries.hpp`,...

```c++ void DBConfig::SetOptionByName(const string &name, const Value &value) { auto option = DBConfig::GetOptionByName(name); if (option) { SetOption(*option, value); return; } auto param = extension_parameters.find(name); if (param != extension_parameters.end()) { Value...

This should mention the differences in NULL behavior between `concat` and `||` I think just listing `concat` as an alias is incorrect

This issue stems from `numpy` not really being a first class citizen, it's more a side-effect of the `pandas` implementation. If we look at Polars, we can see that for...

Hmm, so that error is thrown here: https://github.com/duckdb/duckdb-iceberg/blob/b45ddb2dd11ffe8765b590f31b0e3f0e09396cda/src/base_manifest_reader.cpp#L29 It's caused by the manifest (list or file) not having the required fields or missing field ids etc.. We should make this...