archie.md

archie.md

@archie.sarrewood.com

SQL, data viz, duckdb, markdown building @evidence.dev archie.evidence.app

2.6K followers245 following425 posts

an awesome write up of the @duckdb-gsheets.com extension

DuckDB@duckdb.org

New blog post by @a13x.bsky.social and @archie.sarrewood.com: Reading and Writing Google Sheets in DuckDB This post introduces the gsheets community extension, which enables DuckDB to directly read from and write to Google Sheets. duckdb.org/2025/02/26/g...

archie.md reposted

We are happy to announce DuckDB v1.2.0 “Histrionicus”! The new release has several usability, security and performance improvements. It has new features for the CSV and Parquet formats, as well new (opt-in) options for DuckDB's own format. Read more at duckdb.org/2025/02/05/a...

@duckdb-gsheets.com v0.0.5 is now live! big news with this version is now you can force duckdb to read EVERYTHING as a string with the all_varchar flag we also welcomed a new contributor to the project in this release - props to Michael Harris for all the new features!

sometimes you get really deep into something very niche i'm always using CSVs + json to store little bits of data, but never considered why these formats (a large part is portability!) Alex and I put together this guide about when to use flat files, and when not to evidence.dev/blog/what-is...

evidence.devWhat Is a Flat File?The ultimate guide to flat files, their use cases, format specifics, and how they stack up against non-flat file databases and relational DBMSs

archie.md reposted

We've overhauled our VS Code autocomplete system to include: - Autocomplete for component props - Inline prop documentation - Option selection for props that have finite options - Query name suggestions within any data prop - Better table suggestions in SQL, including CTEs and chained queries

was interested to see if I could extend this to create a deterministic way to choose a chart based on your data: - how many dimensions? - discrete or continuous? - how many values? etc this is my first attempt - what did i get wrong here?

archie.md@archie.sarrewood.com

i tried to write useful descriptions about when you should choose each chart type in the @evidence.dev library docs.evidence.dev/components/charts/line-chart

also when you are building charts for BI, there is not always a specific predetermined goal for an charts, so much as the more generic goal: track this thing this is different for ad-hoc analysis where there is a predefined question

was interested to see if I could extend this to create a deterministic way to choose a chart based on your data: - how many dimensions? - discrete or continuous? - how many values? etc this is my first attempt - what did i get wrong here?

archie.md@archie.sarrewood.com

i tried to write useful descriptions about when you should choose each chart type in the @evidence.dev library docs.evidence.dev/components/charts/line-chart

having dinner with a friend: “so @evidence.dev just extends markdown to run SQL?” about to tell him this is over simplistic… …realise simpler is better, so just reply: “yeah” friend: “smart” stealing this as my new elevator pitch

proud of myself for tracking down a gnarly bug by learning `git log --online --graph` it's not often that the github UI fails me, but if you are looking for a bug in a large feature branch, then its chronological commit view can't help you (and wont show >250 commits in a PR)

the reason the graph view is so helpful is if you have merge commits in your branch, and you are trying to find two parents, sometimes the second parent is along time ago

linkedin seriously feels like it’s serving me someone else’s feed 10 notifications in a row about comments by a single connection on posts by people i’ve never met in under an hour

not one american I asked in the last week knew that “die hard” is german for “the tower” they literally thought it was english like are you kidding me? the bad guy is called hans gruber and it’s set in a tower block

so much cool stuff coming to Evidence

Evidence@evidence.dev

We're excited to publicly share the Evidence Roadmap! - Custom Themes & Dark Mode (Shipped!) - Pivot Table & SQL Console - Cross Filters & Sharable Links - Prop Autocomplete - Deploy Previews & Email Notifications - Multi Tenant Config & RLS - UI Revamp & Scheduled Emails - EU region for Cloud