That blog post was well written and clear. I love the simplicity of just having the library and a DB provide durable execution guarantees. Really simplifies things. Could be useful in some of my data eng projects.

I wonder if GCP / BigQuery will use this as a forcing function to improve their managed iceberg offering? Exciting year in data.

Thanks for sharing! It will be interesting to see how much support exists for flexible partitioning (sharding) and clustering (sorting) schemes — to truly make these tables performant enough on larger datasets.

This is cool. I wonder if there’s a way to put a comment box directly on the blog post and use the user session (logon info) to the BlueSky app to reduce friction 🤔?

I wrote about this because I often need to be reminded of it. "How to use your impostor syndrome to learn anything":
davidasboth.com/impostor-syn...

I was not aware of that group! I hope to join y’all at a future one.

I was not very active the past year on X, but looking to use this going forward more frequently! Into big data, data engineering, and just a generalist at heart.

Great bokeh on that first one!

👋 Trying to be a more active participant. Really a jack of all trades personally and in the data eng space professionally.

I can’t speak to specific use cases but we’re using it. Massive (sorted) tables are generated; we update the catalog pointer and serve those datasets to end users, avoiding any “load” into a data store. The current metadata for Iceberg serves as an efficient “index” for range and point queries.

Looking forward to hearing more about how CB will support this open table format. Indexes in Postgres or via the Puffin spec would be compelling, IMO.

Good write up indeed! I appreciated hearing about the gradual journey from a single service with Postgres to true federation. Lots to learn from that and article.