这个答案不使用像readdirSync或statSync这样的阻塞函数。它不使用外部依赖关系，也不陷入回调地狱的深渊。

相反，我们使用现代JavaScript的便利，如Promises和async-await语法。异步结果是并行处理的;〇不是按顺序

const { readdir, stat } =
  require ("fs") .promises

const { join } =
  require ("path")

const dirs = async (path = ".") =>
  (await stat (path)) .isDirectory ()
    ? Promise
        .all
          ( (await readdir (path))
              .map (p => dirs (join (path, p)))
          )
        .then
          ( results =>
              [] .concat (path, ...results)
          )
    : []

我将安装一个示例包，然后测试我们的函数-

$ npm install ramda
$ node

让我们看看它是如何工作的

> dirs (".") .then (console.log, console.error)

[ '.'
, 'node_modules'
, 'node_modules/ramda'
, 'node_modules/ramda/dist'
, 'node_modules/ramda/es'
, 'node_modules/ramda/es/internal'
, 'node_modules/ramda/src'
, 'node_modules/ramda/src/internal'
]

使用一个通用模块Parallel，我们可以简化dirs -的定义

const Parallel =
  require ("./Parallel")

const dirs = async (path = ".") =>
  (await stat (path)) .isDirectory ()
    ? Parallel (readdir (path))
        .flatMap (f => dirs (join (path, f)))
        .then (results => [ path, ...results ])
    : []

上面使用的Parallel模块是从一组旨在解决类似问题的函数中提取出来的模式。更多解释请参见相关问答。

你可以使用graph-fs

const {Node} = require("graph-fs");
const directory = new Node("/path/to/directory");

const subDirectories = directory.children.filter(child => child.is.directory);

以防其他人从网络搜索到这里，并且已经在他们的依赖列表中有Grunt，这个问题的答案变得微不足道。以下是我的解决方案:

/**
 * Return all the subfolders of this path
 * @param {String} parentFolderPath - valid folder path
 * @param {String} glob ['/*'] - optional glob so you can do recursive if you want
 * @returns {String[]} subfolder paths
 */
getSubfolders = (parentFolderPath, glob = '/*') => {
    return grunt.file.expand({filter: 'isDirectory'}, parentFolderPath + glob);
}

承诺

import { readdir } from 'fs/promises'

const getDirectories = async source =>
  (await readdir(source, { withFileTypes: true }))
    .filter(dirent => dirent.isDirectory())
    .map(dirent => dirent.name)

回调

import { readdir } from 'fs'

const getDirectories = (source, callback) =>
  readdir(source, { withFileTypes: true }, (err, files) => {
    if (err) {
      callback(err)
    } else {
      callback(
        files
          .filter(dirent => dirent.isDirectory())
          .map(dirent => dirent.name)
      )
    }
  })

Syncronous

import { readdirSync } from 'fs'

const getDirectories = source =>
  readdirSync(source, { withFileTypes: true })
    .filter(dirent => dirent.isDirectory())
    .map(dirent => dirent.name)

另一种递归方法

感谢Mayur了解我的withFileTypes。我写了下面的代码来递归地获取特定文件夹的文件。可以很容易地修改它以只获取目录。

const getFiles = (dir, base = '') => readdirSync(dir, {withFileTypes: true}).reduce((files, file) => {
    const filePath = path.join(dir, file.name)
    const relativePath = path.join(base, file.name)
    if(file.isDirectory()) {
        return files.concat(getFiles(filePath, relativePath))
    } else if(file.isFile()) {
        file.__fullPath = filePath
        file.__relateivePath = relativePath
        return files.concat(file)
    }
}, [])

你可以使用dree，如果使用一个模块是负担得起的

const dree = require('dree');

const options = {
  depth: 1
};
const fileCallback = function() {};

const directories = [];
const dirCallback = function(dir) {
 directories.push(dir.name);
};

dree.scan('./dir', {});

console.log(directories);

指定路径("./dir")的子目录将被打印。

如果您不设置选项depth: 1，您甚至会以递归的方式获取所有目录，而不仅仅是指定路径的有向子目录。