从节点中已解析的csv文件构建对象数组

Phi*_*ien 14 javascript csv arrays object node.js

我有多个表单的csv文件

  • model1A
  • model1B
  • model2A
  • model2B

其中每个csv是一个数组即 model1A = [1, 1, 1]

我想解析这些csv并创建一个包含所有这些模型的数组,其中数组中的每个元素都是对应于一个特定模型的对象,即

finalArray = [ 
  { 
    "model" :   "model1",
    "A"     :   [1, 1, 1],
    "B"     :   [2, 2, 2]
  },
  { 
    "model" :   "model2",
    "A"     :   [3, 3, 3],
    "B"     :   [4, 4, 4]
  }
]
Run Code Online (Sandbox Code Playgroud)

我到目前为止的代码是

var csv = require('csv');
var fs = require('fs');
var parser = csv.parse();
var util = require('util');
var junk = require('junk');
var _ = require('lodash');
var models = [];


fs.readdir(__dirname+'/data', function(err, files) {
    var model = {};
    _.forEach(files, function(n, key) {

        console.log('Analysing file: ' + n);
        var modelName;
        var modelNum;
        var modelParam;


        modelNum = n.match(/\d+/)[0];
        modelName = 'model' + modelNum;
        modelParam = (n.substring(0, n.indexOf('.'))).replace(modelName,'');

        model.model = modelName;
        model[modelParam] = [];
        models.push(model);

        //if (Object.keys(model).length === 3) {
        //    models.push(model);
        //    model = {};
        //}


        fs.createReadStream(__dirname+'/data/'+n).pipe(csv.parse()).pipe(csv.transform(function(row) {
            model[modelParam].push(row);

        })).on('readable', function(){
            while(this.read()){}
        }).on('end', function() {
            console.log('finished reading file ' + n);
            if (key === (files.length - 1)) {
                fs.writeFile('result.json', JSON.stringify(models), function (err) {
                    if (err) throw err;
                    console.log(models.length + ' model(s) parsed');
                    console.log('done');
                });
            }

        }).on('error', function(error) {
            console.log(error);
        });    
    });
});
Run Code Online (Sandbox Code Playgroud)

我知道我的一个问题是我很快将模型推送到数组,导致下面的表格的最后一个数组,其中model1被覆盖model2

[ { model: 'model2', A: [], B: [] },
  { model: 'model2', A: [], B: [] },
  { model: 'model2', A: [], B: [] },
  { model: 'model2', A: [], B: [] } ]
Run Code Online (Sandbox Code Playgroud)

这就是我尝试这段代码的原因

if (Object.keys(model).length === 3) {
  models.push(model);
  model = {};
}
Run Code Online (Sandbox Code Playgroud)

但当然这不起作用,因为它fs.createReadStream是异步的,我正在清理模型,model = {}然后才能正常运行.

我现在正处于舞台上,我觉得自己正在四处乱逛,只是让事情变得更糟.我想创建一些更通用的东西,但是,现在我很高兴让它适用于这里提供的案例,然后我可以看看它是否有所改进.

任何帮助将非常感激!


更新1

继saquib khan关于移动var model = {}内部循环的建议帮助我更接近目标,但它仍然不对.以下是当前结果

[
    {
        "model": "model1",
        "A": [
            [
                "1"
            ],
            [
                "2"
            ],
            [
                "3"
            ],
            [
                "4"
            ]
        ]
    },
    {
        "model": "model1",
        "B": [
            [
                "1"
            ],
            [
                "2"
            ],
            [
                "3"
            ],
            [
                "4"
            ]
        ]
    },
    {
        "model": "model2",
        "A": [
            [
                "1"
            ],
            [
                "2"
            ],
            [
                "3"
            ],
            [
                "4"
            ]
        ]
    },
    {
        "model": "model2",
        "B": [
            [
                "1"
            ],
            [
                "2"
            ],
            [
                "3"
            ],
            [
                "4"
            ]
        ]
    }
]
Run Code Online (Sandbox Code Playgroud)

更新2

根据Denys Denysiuk的建议,结果更接近我想要的,但仍然很短

[
    {
        "model": "model1",
        "A": [
            "1",
            "2",
            "3",
            "4"
        ]
    },
    {
        "model": "model1",
        "B": [
            "1",
            "2",
            "3",
            "4"
        ]
    },
    {
        "model": "model2",
        "A": [
            "1",
            "2",
            "3",
            "4"
        ]
    },
    {
        "model": "model2",
        "B": [
            "1",
            "2",
            "3",
            "4"
        ]
    }
]
Run Code Online (Sandbox Code Playgroud)

这可以工作,如果我可以以某种方式迭代最终的对象数组,合并具有匹配model名称的对象.我目前正在查看lodash文档,看看我是否可以解决问题.如果我这样做,我会在这里回复.

小智 2

试试这个:

fs.readdir(__dirname+'/data', function(err, files) {

    _.forEach(files, function(n, key) {

        console.log('Analysing file: ' + n);            

        var modelNum = n.match(/\d+/)[0];
        var modelName = 'model' + modelNum;
        var modelParam = (n.substring(0, n.indexOf('.'))).replace(modelName,'');

        var model = {};
        var isNewModel = true;
        for(var i = 0; i < models.length; i++) {
            if(models[i].model == modelName) {
               model = models[i];
               isNewModel = false;
               break;
            }
        }
        if(isNewModel) {
            model.model = modelName;
            models.push(model);
        }

        model[modelParam] = [];

        fs.createReadStream(__dirname+'/data/'+n).pipe(csv.parse()).pipe(csv.transform(function(row) {
            model[modelParam].push(row[0]);

        })).on('readable', function(){
            while(this.read()){}
        }).on('end', function() {
            console.log('finished reading file ' + n);
            if (key === (files.length - 1)) {
                fs.writeFile('result.json', JSON.stringify(models), function (err) {
                    if (err) throw err;
                    console.log(models.length + ' model(s) parsed');
                    console.log('done');
                });
            }

        }).on('error', function(error) {
            console.log(error);
        });    
    });
Run Code Online (Sandbox Code Playgroud)