Support importing data from CSV files without header rows

[Standing-Man]

What problem does the new feature solve?

Currently, importing CSV files into GreptimeDB requires the file to contain a header row.

Supporting CSV imports without header rows would improve usability and flexibility. Users could directly import such files by mapping columns based on the table schema order or by specifying column mappings explicitly.

What does the feature do?

no_header.csv

a1,b1,2026-01-12 00:00:01
a2,b2,2026-01-12 00:00:02
a3,b3,2026-01-12 00:00:03
a4,b4,2026-01-12 00:00:04

create table no_header(a STRING, b STRING, ts TIMESTAMP TIME INDEX, PRIMARY KEY(a));

copy no_header from 'no_header.csv' WITH(FORMAT='csv', has_header='false');

Implementation challenges

No response

[killme2008]

@Standing-Man Good idea!

We could ask users to create the table before importing data.