Hacker News new | ask | show | jobs
by orthoxerox 62 days ago
Where did you find these numbers? The way I see it, it's basically Iceberg, but with all metadata and small data changes stored in an RDBMS.
1 comments
cbility 60 days ago
It depends what you're measuring. For streaming workloads they're reporting 923x speedup for aggregations, and 105x for insertions, over iceberg [1].

There are also some improvements other areas, e.g. cross table transactions are not possible in iceberg/delta, but they are in ducklake. [2]

[1] https://ducklake.select/2026/04/02/data-inlining-in-ducklake...

[2] https://ducklake.select/manifesto/#features

link
orthoxerox 60 days ago
But that's a single use case. I also can put a Postgres DB in front of Iceberg, add some views in Trino to consume the data and get faster streaming ingestion. I know this sounds like that Dropbox is just rsync+cron post, but as soon as you replace pure insertions with upserts the gap should vanish.
link