Feature Request / Improvement
As in the title, right now AddFiles will read the files I just wrote. I'd like to be able to skip the arrow cost (I write from my internal format directly to parquet, while also bookkeeping schema, etc), but if I use the AddFiles API then I immediately am forced to read the files I just wrote, but I'd rather supply the metadata correctly (I'm also setting the field IDs correct which is part of the requirements for reading)
iceberg slack ref: https://apache-iceberg.slack.com/archives/C05J3MJ42BD/p1770135889569909
Feature Request / Improvement
As in the title, right now
AddFileswill read the files I just wrote. I'd like to be able to skip the arrow cost (I write from my internal format directly to parquet, while also bookkeeping schema, etc), but if I use the AddFiles API then I immediately am forced to read the files I just wrote, but I'd rather supply the metadata correctly (I'm also setting the field IDs correct which is part of the requirements for reading)iceberg slack ref: https://apache-iceberg.slack.com/archives/C05J3MJ42BD/p1770135889569909