提高重复查找器和 For 循环性能

Question

我制作了一个脚本，可以在对象数组中找到一个键的重复项。如果找到重复项，则将新键和值（“Duplicate”：true）添加到具有重复键值的对象中。

数据示例

{
    "Id": "1",
    "NI Number": "NG111111A",
    "Full Name": "Test Test Tester",
    "Address Line 1": "My House",
    "Address Line 2": "My Road",
    "Address Line 3": "My Suburb",
    "City / Town": "My Town",
    "Country": "United Kingdom",
    "PostCode": "",
    "Creation Date": "24 December 2014"
},
{
    "Id": "2",
    "NI Number": "NM123405C",
    "Full Name": "A Dummy",
    "Address Line 1": "Dummy 1",
    "Address Line 2": "Dummy 2",
    "Address Line 3": "Dummy 3",
    "City / Town": "Dummy 4",
    "Country": "United Kingdom",
    "PostCode": "",
    "Creation Date": "09 February 2015"
}

脚本

for (let i = 0, len = cleanedData.length; i < len; i++) {

    let foundDuplicate = false;

    if (cleanedData[i]["Duplicate"] === "false" || cleanedData[i]["Duplicate"] === undefined) {

        for (let t = i + 1, len = cleanedData.length; t < len; t++) {

            if (cleanedData[i]["NI Number"] === cleanedData[t]["NI Number"]) {
                foundDuplicate = true;
                cleanedData[t]["Duplicate"] = true;
            }

        }

        if (foundDuplicate === true) {
            cleanedData[i]["Duplicate"] = true;
        } else {
             cleanedData[i]["Duplicate"] = false;
        }
    }

}

我正在尝试在 33,000 条记录中查找重复的“NI 编号”。 NI 编号可以重复多次。该脚本当前按预期工作，但运行时间超过 70 秒。如果可能的话，我想将其缩短到 35 秒。

我是 JavaScript 新手，但从我所阅读的内容来看，使用缓存长度的 for 循环是迭代数组的快速方法。我已阅读该地图，设置可以提高性能，但我不确定如何将它们实现到我的脚本中。

有什么方法可以提高我的代码的性能吗？

Answer 1

您尚未提供任何基准测试工具，因此希望以下内容有所帮助。

您可以减少测试次数，并尽可能少地进行查找。此外，当您分配布尔值但针对字符串进行测试时，您的一些测试是错误的。希望 cmets 足够了。

// Reduce test data to minimum required, increase sample size
var cleanedData = [
  {"Id": "1","NI Number": "NG111111A"},
  {"Id": "2","NI Number": "NM123405C"}, // Duplicate
  {"Id": "3","NI Number": "NM123405D"},
  {"Id": "4","NI Number": "NM123405E"}, // Duplicate
  {"Id": "5","NI Number": "NM123405C"}, // Duplicate
  {"Id": "4","NI Number": "NM123405E"}, // Duplicate
  {"Id": "4","NI Number": "NM123405F"}, 
  {"Id": "4","NI Number": "NM123405E"}  // Duplicate
 ];

// Use var for better compatibility unless you really need let
for (var i = 0, iLen = cleanedData.length; i < iLen; i++) {
   
  // Store ref to current object for performance
  var obj = cleanedData[i];

  // If doesn't have Duplicate property, add and set to false
  // Only test if not already marked as a duplicate
  if (!('Duplicate' in obj)) {
    obj.Duplicate = false;

    // Reuse iLen, don't need to get length of array again
    for (var t = i + 1; t < iLen; t++) {

      // Store ref to test obj for performance
      var tObj = cleanedData[t];

      // Simplify assignment to both objects if duplicate found
      if (obj['NI Number'] === tObj['NI Number']) {
        obj.Duplicate = true;
        tObj.Duplicate = true;
      }
    }
  }
}

console.log(cleanedData)

如果它有助于提高性能，请告诉我。您也可以使用 forEach 之类的数组方法，但我认为 for 循环在某些主机中至少同样快而且快得多。

按照 Salman A 的建议实施索引将类似于以下内容，它简洁（嗯，比第一种方法少一行）并且可能非常快，因为它只循环一次，但它确实对 索引。

var cleanedData = [
  {"Id": "1","NI Number": "NG111111A"},
  {"Id": "2","NI Number": "NM123405C"}, // Duplicate
  {"Id": "3","NI Number": "NM123405D"},
  {"Id": "4","NI Number": "NM123405E"}, // Duplicate
  {"Id": "5","NI Number": "NM123405C"}, // Duplicate
  {"Id": "4","NI Number": "NM123405E"}, // Duplicate
  {"Id": "4","NI Number": "NM123405F"}, 
  {"Id": "4","NI Number": "NM123405E"}  // Duplicate
 ];
 
// Store NI Number value and index first found
var index = {};

for (var i=0, iLen=cleanedData.length; i<iLen; i++) {
  var obj = cleanedData[i];
  var value = obj['NI Number'];

  // If obj doesn't have Duplicate property, add as false
  if (!('Duplicate' in obj)) obj.Duplicate = false;

  // If value is in index, mark duplicates
  if (value in index) {
    obj.Duplicate = true;
    cleanedData[index[value]].Duplicate = true;
    
   // Otherwise, add value to index
  } else {
    index[value] = i;
  }
}

console.log(cleanedData);