PowerShell中如何将JSON字符串转为JSON对象存入JSON数组？

问题描述 #

编写PowerShell脚本读取JSON文件时，通过Get-Content -Path $_.FullName将文件内容以字符串形式读取并存储到JSON数组中，但希望Contents字段是JSON对象而非JSON字符串的转义表示。

当前输出 #

[
  {
    "LastWriteTime": "2023-03-07T15:22:07.8269947-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-07-202100Z_1678220527823.json",
    "Contents": "{\"a\":12345,\"b\":\"This is a test\\nof wills.\\n\",\"c\":[\"ABC006\",\"ABC502\",\"ABC503\"]}"
  },
  {
    "LastWriteTime": "2023-03-08T15:18:58.4251682-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-08-201800Z_1678306738418.json",
    "Contents": "{\"a\":23456,\"b\":\"To\\nstack overflow\\n\",\"c\":[\"ABC006\",\"ABC502\",\"ABC503\"]}"
  },
  {
    "LastWriteTime": "2023-03-09T15:06:18.6890386-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-09-200600Z_1678392378684.json",
    "Contents": "{\"a\":34567,\"b\":\"or not to\\nstack overflow\\nthat is the question.\\n\",\"c\":[\"ABC006\",\"ABC502\",\"ABC503\"]}"
  }
]

期望输出 #

[
  {
    "LastWriteTime": "2023-03-07T15:22:07.8269947-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-07-202100Z_1678220527823.json",
    "Contents": {"a":12345,"b":"This is a test\nof wills.\n","c":["ABC006","ABC502","ABC503"]}
  },
  {
    "LastWriteTime": "2023-03-08T15:18:58.4251682-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-08-201800Z_1678306738418.json",
    "Contents": {"a":23456,"b":"To\nstack overflow\n","c":["ABC006","ABC502","ABC503"]}
  },
  {
    "LastWriteTime": "2023-03-09T15:06:18.6890386-05:00",
    "FullName": "\\server\d$\Program\logs\alerts\2480702_2023-03-09-200600Z_1678392378684.json",
    "Contents": {"a":34567,"b":"or not to\nstack overflow\nthat is the question.\n","c":["ABC006","ABC502","ABC503"]}
  }
]

原脚本片段 #

$rx = -join ($id, "_.*\.json")
$dir = -join ("\\", $machine, "\d$\\", $appname, "\logs\alerts")
$answer = [System.Collections.ArrayList]@()
foreach ($machine in $machines)
{
    $arr = Get-ChildItem $dir -Recurse |
      Where-Object { $_.Name -match $rx } |
      Select-Object @{ n = 'LastWriteTime'; e = { $_.LastWriteTime } },
        @{ n = 'FullName'; e = { $_.FullName } },
        @{ n = 'Contents'; e = { Get-Content -Path $_.FullName } }
}

$null = $answer.Add($arr)
$answer | Sort-Object { $_.LastWriteTime -as [datetime] }

解决方案 #

核心是将读取到的JSON字符串转换为PowerShell对象，这样在最终序列化输出时，该对象会被转为JSON对象而非转义字符串。同时修正原脚本中的路径定义和集合添加逻辑，避免错误：

$rx = -join ($id, "_.*\.json")
$answer = [System.Collections.ArrayList]@()
foreach ($machine in $machines)
{
    # 将路径定义移至foreach内部，确保$machine变量已赋值
    $dir = -join ("\\", $machine, "\d$\", $appname, "\logs\alerts")
    $arr = Get-ChildItem $dir -Recurse |
      Where-Object { $_.Name -match $rx } |
      Select-Object @{ n = 'LastWriteTime'; e = { $_.LastWriteTime } },
        @{ n = 'FullName'; e = { $_.FullName } },
        # 读取完整字符串并转为PowerShell对象
        @{ n = 'Contents'; e = { Get-Content -Path $_.FullName -Raw | ConvertFrom-Json } }
    
    # 用AddRange添加集合元素，避免嵌套数组
    $null = $answer.AddRange($arr)
}

# 排序后转为JSON，指定深度避免结构截断
$answer | Sort-Object { $_.LastWriteTime -as [datetime] } | ConvertTo-Json -Depth 10

关键说明 #

• Get-Content -Raw：读取整个文件为单一字符串，避免默认按行返回字符串数组导致的解析失败；
• ConvertFrom-Json：将JSON字符串转为PowerShell自定义对象，后续序列化时会自动转为原生JSON对象；
• AddRange：替代原脚本的Add方法，避免将整个集合作为单个元素存入列表；
• -Depth 10：解决ConvertTo-Json默认深度不足导致的嵌套结构截断问题，可根据实际JSON结构调整数值。

内容的提问来源于stack exchange，提问作者gknauth