Commander.js: how to use sub command with option?

Created on 7 Apr 2016 · 13Comments · Source: tj/commander.js

Hi I am trying sub command with option. Is there an example? I tried below code. How can I get the option value from the action callback?

  .command('upload')
  .description('upload a file')
  .option('-f, --file <file>', 'file name')
  .action(function(file) {
    console.log("uploading...");
    console.log(file);
  });

 program.parse(process.argv);

$ node index.js upload -f filename
uploading...
Command {
commands: [],
options:
[ Option {
flags: '-f, --file ',
required: -12,
optional: 0,
bool: true,
short: '-f',
long: '--file',

Source

kylelix7

👍11

Most helpful comment

If you want git-style sub commands this is something I would do:

/var/foo/bar/main:

#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('create', 'command description')
    .parse(process.argv);

/var/foo/bar/main-create:

#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('cluster','command description')
    .parse(process.argv);

/var/foo/bar/main-create-cluster

#!/usr/bin/env node
'use strict';
const program = require('commander');
program
    .version('0.0.1')
    .option('-n, --name <name>', 'Sets the cluster name')
    .option('-i, --ip <ip>', 'Sets the cluster ip')
    .parse(process.argv);

but note that commander won't show you help usage for options.

gentunian on 24 Sep 2016

👍13 🚀1 😕1

All 13 comments

Hi @kylelix7, the last argument passed to the callback is actually an options object so you should be able to get the file name value with options.file.

  .command('upload')
  .description('upload a file')
  .option('-f, --file <file>', 'file name')
  .action(function(options) {
    console.log("uploading...");
    console.log(options.file);
  });

program.parse(process.argv);

If you have arguments available to the command (as in program.command('upload' [fileName])), they will be expected arguments in the callback function. See https://www.npmjs.com/package/commander#examples. Hope this helps!

drewbrokke on 7 Apr 2016

What if you're using git-style commands in different files? The options don't seem to come through when chaining the .action cb

cameronroe on 18 Aug 2016

If you want git-style sub commands this is something I would do:

/var/foo/bar/main:

#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('create', 'command description')
    .parse(process.argv);

/var/foo/bar/main-create:

#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('cluster','command description')
    .parse(process.argv);

/var/foo/bar/main-create-cluster

#!/usr/bin/env node
'use strict';
const program = require('commander');
program
    .version('0.0.1')
    .option('-n, --name <name>', 'Sets the cluster name')
    .option('-i, --ip <ip>', 'Sets the cluster ip')
    .parse(process.argv);

but note that commander won't show you help usage for options.

gentunian on 24 Sep 2016

👍13 🚀1 😕1

Have a related question about subcommands... what if want to use common options across subcommands. Based on the above it seems I have to define them in all my subcommands. Am I missing something?

For example, I want to have a -l --loglevel [debug|info|warn|error] option for all my commands & subcommands. It seems I need to set a program.option in each top level command and all subcommands.

I guess I could have a function that takes program as a param, or is there a better way?

You can see my current attempts in https://github/dgem/kat-nip

dgem on 6 Feb 2017

👍8

@gentunian Hey what if at the first level, I want to list the available sub-commands of 'create'? For example right now if I do
$cli-name create
by itself it will raise error since I don't give any sub-command.

frankcchen on 19 Oct 2018

@frankcchen I cannot reproduce that:

~/code/nodejs/borrar/main.js:

#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('create', 'command description')
    .parse(process.argv);

~/code/nodejs/borrar/main-create.js:

```javascript
#!/usr/bin/env node
'use strict';
const program = require('commander');

program
    .version('0.0.1')
    .command('cluster','command description')
    .parse(process.argv);

Test:

[seba@leyla:~/code/nodejs/borrar] 2s $ node main.js 
Usage: main [options] [command]

Options:
  -V, --version  output the version number
  -h, --help     output usage information

Commands:
  create         command description
  help [cmd]     display help for [cmd]
[seba@leyla:~/code/nodejs/borrar] $ node main.js create
Usage: main-create [options] [command]

Options:
  -V, --version  output the version number
  -h, --help     output usage information

Commands:
  cluster        command description
  help [cmd]     display help for [cmd]
[seba@leyla:~/code/nodejs/borrar] $ node main-create.js 
Usage: main-create [options] [command]

Options:
  -V, --version  output the version number
  -h, --help     output usage information

Commands:
  cluster        command description
  help [cmd]     display help for [cmd]

gentunian on 19 Oct 2018

@gentunian You're right.
What if I want to handle unsupported command at the top level aka main.js? Assume your program only has command 'create'. now if I do
$ node main.js bogus
It will have no output.

I have tried adding
.on('command:* , () => {...});
But it didn't really work. The added line intercepts all commands basically.

frankcchen on 23 Oct 2018

@frankcchen I think you could try and play with the result of parsing the options/commands.

Play with this:

/var/foo/main.js:

#!/usr/bin/env node
'use strict';
const program = require('commander');

const r = program
    .version('0.0.1')
    .command('create', 'command description')
    .parse(process.argv);

if (r) {
   console.log('I think no matches were found');
}

I've edited the code. I think you could do a workaround just like that. Each sub-command may also follow the same workaround.

For reference what parse() function returns check the source code

gentunian on 23 Oct 2018

I understand the process of how Commander handles subcommands and how each level is a js file in its own right - although i think this is a very cludgy approach.
I will have around 15 first level commands and maybe as many 10 second level commands for each of the first level command followed by optional parameters - so that is going to lead to a huge number of sub files all in the same top level folder - just to manage the process.
for example:
job list -prefix
job status -jname
job purge -jname
job submit -batchfile

in this example job is a top level command category and the second command (list etc) is the action to take within that category. I would like to isolate all the subcommand files relating to jobs into a jobs folder. That way I would at least be able to manage the 15 or so sub folders for my top level commands.

Rather than be fixed at using level 1 file followed by level 2 file having to be in the same folder is
there any means of being able to specify the structure as to where these files can be found?
I am at the point of wondering if Commander is actually the best tool for me to use for the type of Command Line Interface that I want to put in place?

What I would like to do seems reasonable but I have found no tool that will help accomplish what I want to do whilst also providing built in help.

davidkjackson54 on 27 Dec 2018

Following seems to work for me - though I'd also appreciate a proper solution. Grouping of commands and options would be nice too. And some option names conflict with the implementation of commander, e.g. 'name' or 'domain'.

    .version('0.1.0')
    .description('CLI');

commander
    .command('deployment <command>')
    .description('Manage deployment')
    .option('')
    .option('create', 'Create deployment')
    .option('--cloud <provider>', 'Cloud provider', 'aws')
    .option('--region <region>', 'Region', 'eu-central-1')
    .option('--prefix <prefix>', 'Resource prefix', 'tld-domain')
    .option('')
    .option('update', 'Update deployment')
    .option('--prefix <prefix>', 'Resource prefix', 'tld-domain')
    .action((command, options) => {
        console.log(command, options);
    });

Can be invoked either way (though the help will produce Usage: deployment [options] <command>):

cli deployment create --prefix=my-domain
cli deployment --prefix=my-domain create

Generated Help - 1st Level

# cli --help
Usage: cli [options] [command]

CLI

Options:
  -V, --version                   output the version number
  -h, --help                      output usage information

Commands:
  deployment [options] <command>  Manage deployment

Generated Help - 2nd Level

# cli deployment --help
Usage: deployment [options] <command>

Manage deployment

Options:

  create              Create deployment
  --cloud <provider>  Cloud provider (default: "aws")
  --region <region>   Region (default: "eu-central-1")
  --prefix <prefix>   Resource prefix (default: "tld-domain")

  update              Update deployment
  --prefix <prefix>   Resource prefix (default: "tld-domain")

evalcodenet on 14 Jan 2019

👍1

There was an answer posted to the original post, which was options for a subcommand with an action handler.

There was an example posted for option getting passed down into sub-sub-command.

563 is an open issue about program level options being available to executable subcommands.

Closing this issue as not covering something that will be actioned. Feel free to open a new issue if it comes up again, with new information and renewed interest.

Thank you for your contributions.