使用 PowerShell 脚本从 Azure Blob 存储读取 JSON 文件并写回 Blob 存储中的另一个文件

Question

I have following JSON file (product.json) stored in Azure Blob storage. Is it possible to write PowerShell script to read this file from blob storage make some changes and write back to another file. The output file I would like where following changes should occur:

1. Replace all "_id" with "id"
2. Remove all "_rev" and their values.

Product.json

[
  {
    "_id": "9f4da9d6babeb9d411c896baa68c94c8",
    "_rev": "1-4259271795225df18768ab68baacc96c",
    "account_id": 692278,
    "limit": 10000,
    "products": [
      "Commodity",
      "InvestmentStock"
    ]
  },
  {
    "_id": "cc4b59f585b8556a2bedca78294a0797",
    "_rev": "1-410e479257faba0457bd9b4816c4dc95",
    "account_id": 328304,
    "limit": 10000,
    "products": [
      "Derivatives",
      "InvestmentStock",
      "CurrencyService"
    ]
  },
  {
    "_id": "d7e2a72963cff2760514ff772969ffe0",
    "_rev": "1-2ec6e2679eae13b76410c93f49c14c4a",
    "account_id": 674364,
    "limit": 10000,
    "products": [
      "InvestmentStock"
    ]
  }
]

The outputfile.json should be as follows:

[
  {
    "id": "9f4da9d6babeb9d411c896baa68c94c8",
    "account_id": 692278,
    "limit": 10000,
    "products": [
      "Commodity",
      "InvestmentStock"
    ]
  },
  {
    "id": "cc4b59f585b8556a2bedca78294a0797",
    "account_id": 328304,
    "limit": 10000,
    "products": [
      "Derivatives",
      "InvestmentStock",
      "CurrencyService"
    ]
  },
  {
    "id": "d7e2a72963cff2760514ff772969ffe0",
    "account_id": 674364,
    "limit": 10000,
    "products": [
      "InvestmentStock"
    ]
  }
]

Answer 1

Not the most elegant way, but this should do the trick:

$accountName = "account-name"
$accountKey = "account-key"
$containerName = "container-name"
$blobName = "Product.json"
$outputBlobName = "Output.json"

$ctx = New-AzStorageContext -StorageAccountName $accountName -StorageAccountKey $accountKey

#Read blob and save it to local file
Get-AzStorageBlobContent -Blob $blobName -Container $containerName -Destination "Product.json" -Context $ctx

#Read local file and get JSON object
$jsonContent = Get-Content -Raw -Path "Product.json" | ConvertFrom-Json

#Loop through json content and manipulate it
For ($i=0; $i -lt $jsonContent.Length; $i++) {
    $jsonContent[$i] | Add-Member -NotePropertyName "id" -NotePropertyValue $jsonContent[$i]._id
    $jsonContent[$i].PsObject.Properties.Remove("_id")
    $jsonContent[$i].PsObject.Properties.Remove("_rev")
}
$jsonContent

#Write content back to local disk
$jsonContent | ConvertTo-Json | Set-Content -Path $outputBlobName

#Upload into Azure Storage
$properties = @{"ContentType" = "application/json"}
Set-AzStorageBlobContent -File $outputBlobName -Container $containerName -Blob $outputBlobName -Properties $properties -Context $ctx

Answer 2

If you want to store the file in memory, you can use the DownloadText() method to download the content into memory.

The sample code:

$accountName = "xxx"
$accountKey = "xxx"
$containerName = "xxx"
$blobName = "Product.json"
$outputBlobName = "Output.json"    

$context = New-AzStorageContext -StorageAccountName $accountName -StorageAccountKey $accountKey
$container_client = Get-AzStorageContainer -Name $containerName -Context $context
$source_blob_client = $container_client.CloudBlobContainer.GetBlockBlobReference($blobName)

#download the blob as text into memory
$download_file = $source_blob_client.DownloadText()

$jsonContent = $download_file | ConvertFrom-Json

#Loop through json content and manipulate it
For ($i=0; $i -lt $jsonContent.Length; $i++) {
    $jsonContent[$i] | Add-Member -NotePropertyName "id" -NotePropertyValue $jsonContent[$i]._id
    $jsonContent[$i].PsObject.Properties.Remove("_id")
    $jsonContent[$i].PsObject.Properties.Remove("_rev")
}

$replaced_json = $jsonContent | ConvertTo-Json

#upload the json file
$dest_blob_client =  $container_client.CloudBlobContainer.GetBlockBlobReference($outputBlobName)
$dest_blob_client.Properties.ContentType = "application/json"
$dest_blob_client.UploadText($replaced_json)

Write-Output("**completed**")