Dan Goldin

Annual New Year's Eve post to make sure your date formatting is correct - don't make the same mistake I did in 2018!

yyyy vs YYYY matters

medium.com/@dangoldin/j...

Start Tracking Your Unit Costs to Avoid Surprise Cloud Bills Two key practices for managing cloud infrastructure costs, with real examples from building and scaling an advertising exchange

blog.twingdata.com/p/start-trac...

Want to avoid cloud infra bill surprises?

1. Start tagging everything
2. Start tracking your unit costs

Full post in comments.

Same Query, Different Hash: An exploration into Snowflake's query hashing blind spots We dug into cases where query_parameterized_hash falls down, so you don't have to

Full blog post: blog.twingdata.com/p/same-query...

Snowflake's query_parameterized_hash is supposed to spot useful query patterns, but as it turns out, their rudimentary logic misses many query structures.

Pretty cool - Cursor released a "bug finder" feature and I tested it out against a PR we had open. Results are pretty good. Top one was an intentional change to see if performance would suffer and the others were false flags. Surprised by how expensive it was though. The $2.38 was for +213/-116

With Amazon's launch of S3 Tables it's pretty obvious where we're headed

Ah interesting. It's now getting to this world that whoever controls the catalog controls the world.

Great writeup! I'm curious how this fits into the multi-data warehouse world. It seems if you want to use Snowflake or Databricks on top you'd need to mirror the catalog, is that right?