I used both Gemini and Claude to query ClickHouse using natural language. I used a simple SQL test found online to try to compare them objectively. I think being able to do this type of generative BI is really becoming viable nowadays.
benjaminwootton.com/insights/cli...
I think ClickHouse has been viable for trad data warehousing for a while now, but this jump in join performance removes one of the main objections.
Traditionally, ClickHouse hasn't been that fast when it comes to joins. However, in recent releases it is now performing 20x faster in a TPC-H benchmark. Using another public benchmark a set of join heavy queries over a 500m row dataset now smokes Databricks and Snowflake.
Feel free to ping me.
Later on, if your app throws a frontend error it can go digging around the database and executing queries to understand what is happening before fixing the app. It's crazy to see it building an app in say React then dropping down to the database layer to sling some SQL for debugging purposes.
Here, I've asked it to build a low code app and just given it a hint at the database table. Gemini CLI logs into the database, locates the relevant table, identifies the fields and then builds the app based on what it finds.
Connecting Gemini CLI or Claude Code to your back end database via an MCP server is a really remarkable development experience.
New article on ClickHouse in Capital Markets, where I show how we can investigate a realistic trading scenario to understand the PnL impact of trades being rejected.
benjaminwootton.com/insights/ana...
I'm increasingly seeing ClickHouse Cloud being compared with Snowflake for DWH/BI use cases. I ran this analysis last year and found that CH had a 50% lower cost for equivalent compute, before even accounting for it's better performance/scalability.
benjaminwootton.com/insights/com...
Not only is it the fastest open source way to do CDC from Postgres, it's also been integrated directly into ClickHouse Cloud, meaning that with a few clicks you can subscribe to a change feed from your Postgres database and begin ingesting streaming updates.