config file on startup

[mahiki]

First, thank you so much for your work on this project! Its looking like it will be my favorite way to connect our databricks instance.

Other connectors like duckdb and trino provide for an init_path configuration, which runs initialization commands at the start of the session. Could we have this for databricks too?

Here is what would be in my initialization script every time I log on:

-- ~/.harlequin_dbx_init.sql

use catalog my_team_catalog;
set time zone 'America/Los_Angeles';
declare prev_ds date default current_date - interval '1' day;
select * from myproject_schema.audit_error_log order by error_date desc limit 20;

[alexmalins]

hey @mahiki - this is doable 👍 I'll follow what the duckdb and sqlite adapters do and look for a .databricksrc file in the user's home directory and run that as the initialization script if the user doesn't specify a path explicitly via the --init-path/-i/--init CLI flag

[mahiki]

Awesome! This would add a lot of user ergonomics.

[alexmalins]

Hey @mahiki - check out v0.5.1 released today. It adds a new feature to run a config file automatically on startup.

In your case you'd run harlequin -a databricks --init-path ~/.harlequin_dbx_init.sql ... to run that start up script.

Or if you put those SQL commands into ~/.databricksrc, harlequin-databricks will automatically run them on launch without having to provide an --init-path CLI arg.