Pass extended file info to parquet scan & extended file info to Avro Scan to avoid unnecessary HEAD requests #268
Conversation
| "kind": "git", | ||
| "repository": "https://github.com/duckdb/vcpkg-duckdb-ports", | ||
| "baseline": "0f9bf648ba1ee29291890a1ca9a49a80bba017eb", | ||
| "baseline": "d3a5de6bd45b5907ed189054dd23eff6ec869e14", |
There was a problem hiding this comment.
This is reverting the registry version:
https://github.com/duckdb/vcpkg-duckdb-ports/commits/main/
0f9bf648ba1ee29291890a1ca9a49a80bba017eb - duckdb/vcpkg-duckdb-ports@0f9bf64
d3aa25a569e0eb35e01c59e3e62c9bc7406f0ed9 - duckdb/vcpkg-duckdb-ports@d3aa25a
This version is older than the one that's currently there, is there a reason for that?
I plan to bump this to add avro writes, I'd like to not run into any unrelated issues due to this PR causing issues with a newer patch version
There was a problem hiding this comment.
I don't really see why this change is here, but I also don't think it matters, so we can merge it regardless
There was a problem hiding this comment.
The new commit is the most recent commit to vcpkg-duckdb-ports
d3a5de6bd45b5907ed189054dd23eff6ec869e14 I don't know where d3aa25 is coming from
requires duckdb/duckdb-avro#28
requires duckdb/duckdb#17698
By passing an extended open file info to the avro scan, we can pass
validate_external_file_cache = falseto the FileReader, which means we won't make unnecessary HEAD requests on secondary reads.