What happens?

A CSV file that was previously parseable under v1.2 is no longer parseable under v1.3 because of an error trying to parse timezones. In addition, the ignore_errors = true flag appears to be ignored.

To Reproduce

Consider the following CSV taken from a much larger file:

day_night,operation_no,flight_date_time
Night,38615452,2022/01/27 11:04:57 PM
Night,38615452,2022/01/27 11:04:57 PM
Night,38615475,2022/01/27 11:09:20 PM

Using the CLI with the following command:

select * from read_csv('sample.csv');

Produces the following error:

Not implemented Error:
Unknown TimeZone 'PM'!

Candidate time zones: "Pacific/Marquesas", "PLT", "PNT", "PRC", "GMT"

Setting ignore_errors = true results in the same error as above being generated:

select * from read_csv('sample.csv', ignore_errors = true);

Setting strict_mode = false results in the same error as above being generated:

select * from read_csv('sample.csv', strict_mode = false);

In DuckDB v1.2, this column was parsed as a VARCHAR and no error was reported. If DuckDB v1.3 is now "correct" it should still honour the ignore_errors options.

This may be related to Sniff Timestamp_TZ from CSV FIles #15730

OS:

macOS

DuckDB Version:

DuckDB v1.3.0 (Ossivalis) 71c5c07 clang-17.0.0

DuckDB Client:

CLI

Hardware:

MacBook Pro

Full Name:

Kenny Carruthers

Affiliation:

None

What is the latest build you tested with? If possible, we recommend testing with the latest nightly build.

I have tested with a stable release

Did you include all relevant data sets for reproducing the issue?

Yes

Did you include all code required to reproduce the issue?

• Yes, I have

Did you include all relevant configuration (e.g., CPU architecture, Python version, Linux distribution) to reproduce the issue?

• Yes, I have