What happens?

I have 14k JSON files, one record per file.
select col1 from '*.json'; returns 7k rows.

select * from read_json_objects('*.json'); returns 14k rows of JSON-as-string.

What's happening? All files contain the column I'm selecting.

To Reproduce

wget https://www.mapledesign.co.uk/weather.tar.gz
tar xzf weather.tar.gz
cd data
duckdb

# This one gets all the rows
select * from read_json_objects('*.json');

# This one gets approx 50% of rows
select "update" from read_json_auto('*.json')

OS:

Ubuntu 18.04 x64

DuckDB Version:

v0.7.1 b00b93f

DuckDB Client:

CLI

Full Name:

Peter Bowyer

Affiliation:

N/A

Have you tried this on the latest master branch?

• I agree

Have you tried the steps to reproduce? Do they include all relevant data and configuration? Does the issue you report still appear there?

• I agree