Project is abandoned, please warn users this is not supported.

[mahiki]

README should have a warning that no issues have been addressed since 2023 and the project is abandoned.

There are other CLI options, please suggest some at the top of the README:

• harlequin
• usql I have not tested this with databricks odbc cnxn, but this shows it does have a dbx driver

[u35253]

You just saved me some time. Thanks for the links, too.

I wonder if the linkage/description from this page should be updated too: https://docs.databricks.com/aws/en/dev-tools/databricks-sql-cli

[u35253]

What are the best replacements for databricks-sql-cli, if it is truly abandoned? Maybe some good options can be mentioned in the comments here.

Regarding Harlequin: not a replacement for databricks-sql-cli

• harlequin

I would state that, in general, Harlequin is definitively not a suitable replacement for databricks-sql-cli at this time (and not for the forseeable future).

• Current: Harlequin does not support STDIN and STDOUT (e.g., basic terminal based usage), but is instead a "TUI/IDE". In other words, there is no -e <statement> or -f <file> option, to submit statements and direct output to STDOUT, bypassing the whole "text IDE" experience, for normal CLI tool use cases (e.g., shell scripting).
• When asked, the creator/maintainer of Harlequin stated that that basic input/output outside of the IDE interface is "a totally different set of features than what I'm focused on".

Regarding: usql: does it actually work with Databricks?

• usql I have not tested this with databricks odbc cnxn, but this shows it does have a dbx driver

AFAICT, usql seems to not connect correctly with Databricks personal access tokens by using the databricks-sql-go DSN format that appears suggested.

• The parsing seems quite mangled and just plain wrong in my testing.
• There are also no real examples of a full usql config for Databricks on the usql or the related dburl Github pages.
• The usql possibility seems promising/hopeful, but requires additional debugging and investigation from where I sit. (At least databricks-sql-cli and harlequin can connect to Databricks and are easily configured using hostname, personal access token, and HTTP path, unlike usql AFAICT so far -- despite what the README pages indicate.)

Again, I appreciate the tips, and I hope that solid options become well documented.

[u35253]

Following up: usql does seem to work with databricks after all

The usql config issue for Databricks mentioned in the immediately preceding comment seems to be a non-issue.

• There was just some nuance to saving configuring the format correctly (it seems that the correct [array syntax](https://github.com/xo/usql/blob/2be774d5fd7c660d8f7477f647cec585a3b111bb/contrib/config.yaml#L32C1-L33C1 was already documented and appears to work).
• Thus, regarding usql as a possible alternative tool, the connection string seems to work, and results are returned.
• Details are in a separate (and resolved) usql issue.

[mahiki]

Good news on usql.

Like you mentioned, the databricks connection works only while using a config file. I find that usql is actively maintained, and matches up well with psql (which I consider an industry standard-setter).

Here is a working example:

# ~/.usql_config.yaml
connections:
  mycorp:
    - databricks
    - token:[email protected]:443/sql/1.0/warehouses/8675309abcdefXYZ


# basic connection
usql --config ~/.usql_config.yaml mycorp

# my preferred data pager, and editor
PAGER=pspg EDITOR=vim usql --config ~/.usql_config.yaml mycorp

You can use the \edit metacommand in usql to edit the last command, and there are other useful REPL commands for saving output to file etc.