如何根据单词作为键分隔符读取文本文件的块？

Question

我有一个这种格式的.txt文件：

Part #368 - XXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Part #369 - XXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Part #370 - XXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

我读了这样的文件：

var lines = fs.readFileSync('file.txt', 'utf-8')
.split('\n')
.filter(Boolean);

因此它返回文件行的数组。如何从＆＃34; Part＆＃34;开始获取文件的块？字符串？

var parts = _.filter(lines,function( s ) { return s.indexOf( 'Part' ) !== -1; });

像这样的东西，而不是以＆＃34; Part＆＃34;开头的字符串。我想要来自＆＃34; Part＆＃34;字符串到下一个＆＃34;部分＆＃34;字符串。

Answer 1

这将创建一个数组行数组。

var parts = _.reduce(lines, function( result, line ) {
    if (line.indexOf('Part') !== -1) result.push([]);
    _.last(result).push(line);
    return result;
}, []);

Answer 2

JSON Stream

根据@Brad's suggestion，这是一个从stream.Transform扩展的类，它将文件分隔为JSON数组流：

DEBUG

使用示例：

const { Transform } = require('stream');

class Delimited extends Transform {
  constructor({ delimiter = /\r?\n/g, encoding = 'utf8' } = {}) {
    super();

    // initialize internal values
    this._delimiter = delimiter instanceof RegExp ? delimiter : new RegExp(delimiter, 'g');
    this._encoding = encoding;
    this._buffer = '';
    this._first = true;
  }

  _transform(chunk, encoding, callback) {
    // convert input encoding into output encoding
    // and append to internal buffer
    if (encoding === 'buffer') {
      this._buffer += chunk.toString(this._encoding);
    } else if (encoding === this._encoding) {
      this._buffer += chunk;
    } else {
      this._buffer += Buffer.from(chunk, encoding).toString(this._encoding);
    }

    let partialJSON = '';

    // check if delimiter is found
    if (this._delimiter.test(this._buffer)) {
      // split internal buffer by delimiter
      let sections = this._buffer.split(this._delimiter);
      // put possibly incomplete section from array back into internal buffer
      this._buffer = sections.pop();
      // add each section to partial json array
      sections.forEach(section => {
        partialJSON += `${this._first ? '[' : ','}${JSON.stringify(section)}`;
        this._first = false;
      });
    }

    // push partial json array to readable stream
    callback(null, partialJSON);
  }

  _flush(callback) {
    // add remaining buffer as last section to json array
    callback(null, `${this._first ? '[' : ','}${JSON.stringify(this._buffer)}]`);
  }
}

Try it online!

字符串流

或者，如果您不想将JSON转移到另一个文件，进程或作为客户端响应，则可以通过将输出流设置为const fs = require('fs'); let stream = fs.createReadStream('file.txt', 'utf8'); let transform = new Delimited({ delimiter: /\n\n(?=Part #\d)/g }); let json = ''; transform.on('data', (chunk) => json += chunk); transform.on('end', () => console.log(JSON.parse(json))); stream.pipe(transform); 来将每个部分作为块发送：

objectMode: true

使用示例：

const { Transform } = require('stream');

class Delimited extends Transform {
  constructor(delimiter = /\r?\n/g) {
    super({ objectMode: true });

    // initialize internal values
    this._delimiter = delimiter instanceof RegExp ? delimiter : new RegExp(delimiter, 'g');
    this._encoding = 'utf8';
    this._buffer = '';
    this._first = true;
  }

  _transform(chunk, encoding, callback) {
    // convert input encoding into output encoding
    // and append to internal buffer
    if (encoding === 'buffer') {
      this._buffer += chunk.toString(this._encoding);
    } else if (encoding === this._encoding) {
      this._buffer += chunk;
    } else {
      this._buffer += Buffer.from(chunk, encoding).toString(this._encoding);
    }

    if (this._delimiter.test(this._buffer)) {
      // split internal buffer by delimiter
      let sections = this._buffer.split(this._delimiter);
      // put possibly incomplete section from array back into internal buffer
      this._buffer = sections.pop();
      // push each section to readable stream in object mode
      sections.forEach(this.push, this);
    }

    callback();
  }

  _flush(callback) {
    // push remaining buffer to readable stream
    callback(null, this._buffer);
  }
}

Try it online!