使用 Powershell 将多行文本文件提取到单行 csv

Question

我有一项任务，但没有一种简单的方法可以将一些数据解析为正确的格式。 我拥有的文本文件采用以下格式

#N Last Name: Joe
#D First Name: Doe
#P Middle Name: A
Some Data:
#C ID Number: (1) 12345
#S Status: (1) Active 

#N Last Name: Jane
#D First Name: Doee
#P Middle Name: 
Some Data:
#C ID Number: (1) 11111
#S Status: (1) Active 
ID Number: (2) 1231
Status: (2) Active

这是我试图使用的代码。

$A = Select-String -Pattern "#N" MYFILE.txt;
$B = Select-String -Pattern "#D" MYFILE.txt;
$C = Select-String -Pattern "#P" MYFILE.txt;
$D = Select-String -Pattern "#C" MYFILE.txt;
$E = Select-String -Pattern "#S" MYFILE.txt;

$wrapper = New-Object PSObject -Property @{ FirstColumn = $A; SecondColumn = $B; ThirdColumn = $C; FourthColumn = $D; FifthColumn = $E }
Export-Csv -InputObject $wrapper -Path .\output.csv -NoTypeInformation

这是我得到的结果

"SecondColumn","ThirdColumn","FifthColumn","FourthColumn","FirstColumn"
"System.Object[]","System.Object[]","System.Object[]","System.Object[]","System.Object[]"

我正在寻找的 output 是； #N,#D,#P,#C,#S

 Joe, Doe, A, 12345, Active
 Jane, Doee, , 11111, Active

非常感谢任何和所有帮助。

Answer 1

这是解析该数据块的另一种方法。 我更改了用户信息，以使发生的事情更加明显。 [咧嘴笑]

它能做什么...

• 创建单个多行字符串以使用
  当准备好实际执行此操作时，将整个#region/#endregion块替换为对Get-Content -Raw的调用。
• 定义用户数据块之间的分隔符
  在这种情况下，它是 2 个换行符 - 一个在最后一个数据行的末尾，一个用于空白行。
• 将单个多行字符串拆分为多个此类字符串
• 遍历生成的文本块
• 初始化用于构建 PSCO 的 $Vars
• 将文本块拆分为文本行
• 过滤掉任何不以#开头，然后是字母和最后一个空格的行
• 遍历剩余的行
• 在每行的第二个字符上运行一个switch
• 当它与其中一个代码字母匹配时，解析该行并设置等效 $Var 的值
• 完成对当前字符串集的迭代
• 构建一个[PSCustomObject]来保存值
• 将其发送到$Result集合
• 完成遍历文本块 [外部foreach ]
• 在屏幕上显示集合
• 将集合保存到 CSV

如果您想从 CSV 中删除引号，请不要。 [ grin ] 如果您必须冒险破坏 CSV，那么您可以使用Get-Content从文件中加载行并将引号替换为空。

编码...

#region >>> fake reading in a text file as a single multiline string
#    in real life, use "Get-Content -Raw"
$InStuff = @'
#N Last Name: ALast
#D First Name: AFirst
#P Middle Name: AMid
Some Data:
#C ID Number: (1) 11111
#S Status: (1) Active 

#N Last Name: BLast
#D First Name: BFirst
#P Middle Name: 
Some Data:
#C ID Number: (1) 22222
#S Status: (1) Active 
ID Number: (2) 1231
Status: (2) Active
'@
#endregion >>> fake reading in a text file as a single multiline string

$BlockDelim = ([System.Environment]::NewLine) * 2

$Result = foreach ($Block in ($InStuff -split $BlockDelim))
    {
    # initialize stuff to $Null
    #    this handles non-matches [such as a missing middle name] 
    $FirstName = $MidName = $LastName = $IdNumber = $Status = $Null

    # the "-match" filters for lines that start with a "#", a single letter, and a space
    foreach ($Line in ($Block -split [System.Environment]::NewLine -match '^#\w '))
        {
        switch ($Line[1])
            {
            'N' {
                $LastName = $Line.Split(':')[-1].Trim()
                break
                }
            'D' {
                $FirstName = $Line.Split(':')[-1].Trim()
                break
                }
            'P' {
                $MidName = $Line.Split(':')[-1].Trim()
                break
                }
            'C' {
                $IdNumber = $Line.Split(':')[-1].Trim().Split(' ')[-1].Trim()
                break
                }
            'S' {
                $Status = $Line.Split(':')[-1].Trim().Split(' ')[-1].Trim()
                break
                }
            } # end >>> switch ($Line[1])
        } # end >>> foreach ($Line in ($Block -split [System.Environment]::NewLine))

        # create a custom object and send it out to the collection
        [PSCustomObject]@{
            FirstName = $FirstName
            LastName = $LastName
            MidName = $MidName
            IdNumber = $IdNumber
            Status = $Status
            }
    } # end >>> foreach ($Block in ($InStuff -split $BlockDelim))

# display on screen
$Result

# send to a CSV file
$Result |
    Export-Csv -LiteralPath "$env:TEMP\Veebster_ParsedResult.csv" -NoTypeInformation

屏幕上的 output...

FirstName : AFirst
LastName  : ALast
MidName   : AMid
IdNumber  : 11111
Status    : Active

FirstName : BFirst
LastName  : BLast
MidName   : 
IdNumber  : 22222
Status    : Active

CSV 文件的内容...

"FirstName","LastName","MidName","IdNumber","Status"
"AFirst","ALast","AMid","11111","Active"
"BFirst","BLast","","22222","Active"

请注意，没有错误检测或错误处理。 [咧嘴笑]

Answer 2

这是另一个建议。 我的建议是使用 ConvertFrom-String。

首先我们将制作一个模板，我在您的示例数据中占了垃圾两行。 我真的希望这是一个错字/副本。

$template = @'
#N Last Name {[string]Last*:last1}
#D First Name: {[string]First:first1}
#P Middle Name: {[string]Middle:A}
#C ID Number: (1) {[int]ID:11111}
#S Status: (1) {[string]Status:status1}

#N Last Name: {[string]Last*:Jane}
#D First Name: {[string]First:Doee}
#P Middle Name: {[string]Middle: \s}
#C ID Number: (1) {[int]ID:11111}
#S Status: (1) {[string]Status:Active}
{!Last*:ID Number: (2) 1231
Status: (2) Active}
'@

现在我们将该模板应用于您的数据。 首先我们将解析一个here-string。

@'
#N Last Name: Joe
#D First Name: Doe
#P Middle Name: A
Some Data:
#C ID Number: (1) 12345
#S Status: (1) Active 

#N Last Name: Jane
#D First Name: Doee
#P Middle Name: 
Some Data:
#C ID Number: (1) 11111
#S Status: (1) Active 
ID Number: (2) 1231
Status: (2) Active
'@ | ConvertFrom-String -TemplateContent $template -OutVariable results

Output

Last   : Joe
First  : Doe
Middle : A
ID     : 12345
Status : Active

Last   : Jane
First  : Doee
ID     : 11111
Status : Active

现在我们可以构建我们的 object 以准备导出。

$results | foreach {
    [pscustomobject]@{
        FirstName = $_.first
        LastName  = $_.last
        MidName   = $_.middle
        IdNumber  = $_.id
        Status    = $_.status
    }
} -OutVariable export

现在我们可以导出它了

$export | Export-Csv -Path .\output.csv -NoTypeInformation

这是 output.csv 中的内容

PS C:\> Get-Content .\output.csv
"FirstName","LastName","MidName","IdNumber","Status"
"Doe","Joe","A","12345","Active"
"Doee","Jane",,"11111","Active"

这是从文件中读取它的相同内容。

$template = @'
#N Last Name {[string]Last*:last1}
#D First Name: {[string]First:first1}
#P Middle Name: {[string]Middle:A}
#C ID Number: (1) {[int]ID:11111}
#S Status: (1) {[string]Status:status1}

#N Last Name: {[string]Last*:Jane}
#D First Name: {[string]First:Doee}
#P Middle Name: {[string]Middle: \s}
#C ID Number: (1) {[int]ID:11111}
#S Status: (1) {[string]Status:Active}
{!Last*:ID Number: (2) 1231
Status: (2) Active}
'@

get-content .\ndpcs.txt | 
    ConvertFrom-String -TemplateContent $template | foreach {
        [pscustomobject]@{
            FirstName = $_.first
            LastName  = $_.last
            MidName   = $_.middle
            IdNumber  = $_.id
            Status    = $_.status
        }
    } | Export-Csv -Path .\output.csv -NoTypeInformation

让我们仔细检查一下我们的 CSV 的内容以确定。

Get-Content .\output.csv
"FirstName","LastName","MidName","IdNumber","Status"
"Doe","Joe","A","12345","Active"
"Doee","Jane",,"11111","Active"

需要注意的几点：如果这之后的数据集具有不同的特征，则需要向模板中添加更多样本。 如果两个额外的行（ID 和状态）不应该在那里，只需删除模板的那部分。

我建议大家在制定逻辑/构建脚本时使用 -outvariable 参数，因为您可以看到 output 并同时分配给变量。

Answer 3

这会起作用，但它像罪一样丑陋。

# Get content from text file
$Txtfile = Get-Content "C:\temp\test.txt" -raw
# Add delimiter to split users
$Delimiter = "

"
$Users = $Txtfile -split $Delimiter

# Create an array list to add data to so it can be exported later.
$collectionVariable = New-Object System.Collections.ArrayList
ForEach($Grouping in $Users) {
    $temp = New-Object System.Object
    $temp | Add-Member -MemberType NoteProperty -Name "#N" -Value (([regex]::match($Grouping, '#N.*').value) -split " ")[-1]
    $temp | Add-Member -MemberType NoteProperty -Name "#D" -Value (([regex]::match($Grouping, '#D.*').value) -split " ")[-1]
    $temp | Add-Member -MemberType NoteProperty -Name "#P" -Value (([regex]::match($Grouping, '#P.*').value) -split " ")[-1]
    $temp | Add-Member -MemberType NoteProperty -Name "#C" -Value (([regex]::match($Grouping, '#C.*').value) -split " ")[-1]
    # This is [-2] due to new line at end of the groupings.
    $temp | Add-Member -MemberType NoteProperty -Name "#S" -Value (([regex]::match($Grouping, '#S.*').value) -split " ")[-2]
    $collectionVariable.Add($temp) | Out-Null   
}