Dumky de Wilde | DuckDB Snippets

Back to Code Snippets

Dumky de Wilde

@dumky

‌

Open the HTML version of an EXPLAIN statementBash

DuckDB can output the plan in different ways, e.g. HTML or JSON. It's hard to easily view this when in the DuckDB CLI, but in non-interactive mode we can pass the result to a HTML file directly and open it.

Execute this Bash

duckdb my_complex_db.duckdb -c "                                                                                                                                                         
EXPLAIN  (ANALYZE, FORMAT html)
INSERT INTO big_data.complex_db
    (key_to_update, values, event_time)
SELECT
    key_to_update, values, event_time
FROM big_data.new_data
WHERE event_time >= '2026-01-01'::date
" > output.html && open output.html
Copy code

Dumky de Wilde

Created 01/08/26

‌

Fetch HackerNews Stories with read_json()

You can read JSON from any endpoint into DuckDB. This uses read_json() to fetch all HackerNews 'Who is hiring' stories and create a table with all stories and the comment IDs. It works by: - creating a list of URLs - reading the list and creating the table - querying the table

Create JSON URLsSQL

-- https://hn.algolia.com/api
with urls as (
  select
    'https://hn.algolia.com/api/v1/search_by_date?tags=author_whoishiring&page=' || page_num || '&query=hiring?&attributesToRetrieve=title,tags,comment_text,created_at,object_id,story_id,story_title,children,story_text' as url
  from unnest(range(0,10)) as t(page_num)
)

select array_to_json(list(url)) from urls
Copy code

Read JSON from API endpointSQL

create or replace table hn_stories as (
select 
  hit.title as story_title,
  hit.story_text,
  hit.created_at,
  hit.objectId,
  hit.children,

from read_json(
 -- your_url_list_from_above --
  ), unnest(hits) as t(hit)
)
Copy code

Dumky de Wilde

Created 12/12/25