CSV 到 JSON 数组列表

编程问答 2022-05-10

问题描述

我的csv就是这样的，只有数字。没有标题。我必须在 javascprit (node.js) 代码中识别每一列的名称。

1364.00,0.15,0.36,-0.13,-3.24,-0.42,-0.15,0.90,0.00,-0.01,0.02,0.26,0.01,-0.04
1374.00,0.30,0.76,-0.25,-3.25,-0.41,0.91,-0.00,-0.04
1384.00,0.45,1.08,-0.35,-3.17,-0.10,1.00,-0.07

节点js代码


    const csvFilePath = "test - Cópia.csv"
    
    const csvtojsonV2=require('csvtojson')
    csvtojsonV2({
        noheader:true,checkType:true,headers:["Time","Yaw","Pitch","Roll","heading","Ax","Ay","Az","Gx","Gy","Gz","Mx","My","Mz"],colParser:{        
            "column1": Int32Array,"column2":Int32Array,"column3":Int32Array,"column4":Int32Array,"column5":Int32Array,"column6":Int32Array,"column7":Int32Array,"column8":Int32Array,"column9":Int32Array,"column10":Int32Array,"column11":Int32Array,"column12":Int32Array,"column13":Int32Array,"column14":Int32Array},})
    .fromFile(csvFilePath)
    .then((json) => {
        console.log(json)  
    })

我用上面的代码得到的结果是这个json：

[
      {
        Time: 1364,Yaw: 0.15,Pitch: 0.36,Roll: -0.13,heading: -3.24,Ax: -0.42,Ay: -0.15,Az: 0.9,Gx: 0,Gy: -0.01,Gz: 0.02,Mx: 0.26,My: 0.01,Mz: -0.04
      },{
        Time: 1374,Yaw: 0.3,Pitch: 0.76,Roll: -0.25,heading: -3.25,Ax: -0.41,Ay: -0.13,Az: 0.91,Gy: -0,{
        Time: 1384,Yaw: 0.45,Pitch: 1.08,Roll: -0.35,heading: -3.17,Ay: -0.1,Az: 1,Gx: -0,Mz: -0.07
      }
    ]

我想要一些不同的，像这样（下面：每列一个数组字典。

{"Time": [1364.0,1374.0,1384.0],"Yaw": [0.15,0.3,0.45],"Pitch": [0.36,1.08],"heading": [-0.13,-0.35],"Roll": [-3.24,-3.17],"Ax": [-0.42,-0.41],"Ay": [-0.15,-0.1],"Az": [0.9,1.0],"Gx": [0.0,0.0,-0.0],"Gy": [-0.01,-0.0,-0.01],"Gz": [0.02,0.02],"Mx": [0.26,0.26],"My": [0.01,0.01],"Mz": [-0.04,-0.04,-0.07]}

解决方法

第一种方法

使用使用 csvtojson 库的代码，您只需添加以下内容：

    const csvFilePath = "test - Cópia.csv"
    
    const csvtojsonV2=require('csvtojson')
    csvtojsonV2({
        noheader:true,checkType:true,headers:["Time","Yaw","Pitch","Roll","Heading","Ax","Ay","Az","Gx","Gy","Gz","Mx","My","Mz"],colParser:{        
            "column1": Int32Array,"column2":Int32Array,"column3":Int32Array,"column4":Int32Array,"column5":Int32Array,"column6":Int32Array,"column7":Int32Array,"column8":Int32Array,"column9":Int32Array,"column10":Int32Array,"column11":Int32Array,"column12":Int32Array,"column13":Int32Array,"column14":Int32Array},})
    .fromFile(csvFilePath)
    .then((json) => {
       const RESULT = json.reduce((acc,v,i) => {
          for (const [key,value] of Object.entries(v)) {
            if (!acc[key]) acc[key] = [];
            acc[key].push(value);
          }
          return acc;
        },{});
       console.log(RESULT)
    })

它所做的是在 acc 变量中累积数据，使用每行的键，这些键是标题，正如您在 acc[key].push(value) 中看到的，if (!acc[key]) acc[key] = [] 是使确保使用 acc 变量（累加器）中的空数组初始化标题；

第二种方法

这不使用任何库：

// Arbitrary data (from you example).
const DATA = `1364.00,0.15,0.36,-0.13,-3.24,-0.42,-0.15,0.90,0.00,-0.01,0.02,0.26,0.01,-0.04
1374.00,0.30,0.76,-0.25,-3.25,-0.41,0.91,-0.00,-0.04
1384.00,0.45,1.08,-0.35,-3.17,-0.10,1.00,-0.07
`.trim().split('\n').map(row => row.split(','));

const HEADERS = ["Time","Mz"];

const RESULT = {};

// Assign every heading to the result.
for (const HEADING of HEADERS) RESULT[HEADING] = [];

DATA.map(row =>
  // Each row → each column,number of columns === number of headers
  row.map((column,i) =>
    RESULT[HEADERS[i]]
    .push(Number(column))
  )
);

console.log(RESULT);

每一步的解释：

  1. 假设CSV的内容用“\n”分隔行，首选的列分隔符是“,”，我设置了一个任意的数据内容；
  2. 创建了一个常量 HEADERS，标题必须与 CSV 数据列的长度相同；
  3. 创建了一个常量 RESULT，这会将解析数据的结果存储到 JSON 中。
  4. 我们用定义的 RESULT 和一个空数组初始化 HEADERS 的键；
  5. 我们从每一行然后每一列映射数据，使用列的索引来引用我们HEADERS数组中的当前标题，因为长度列数应等于 HEADERS 常量的长度。

如果您必须保留数字的尾随零（例如：15.00），那么您可以在插入结果数组之前从 Number() 方法中的 Number(column) 中删除 .push() 方法.

csvcsvcsvcsvtojsonjsonpandaspandas