通过regex从行中提取特定的多个字符串

Question

我一直试图用.txt从PowerShell文件中的多行中提取某些值。我有一个很大的文件和所有备份，并试图提取所有这些行。

Txt文件：

Backup-ID:           hostname01
Policy:              VM_weekly
Primary Copy:        23
Expires:             1/5/2024 3:19:13 AM
Type:                4


Copy Number:        2
Fragment Size (KB): 6188832
Expires:            1/5/2024 3:19:13 AM
MediaID:            XXX122
TestID:             1222
Block:              33


Copy Number:        3
Fragment Size (KB): 6188832
Expires:            1/5/2024 3:19:13 AM
MediaID:            XXX134
TestID:             223
Block:              22
Duplicate:          N



Backup-ID:           hostname02
Policy:              VM_weekly2
Primary Copy:        24
Expires:             1/5/2024 3:19:13 AM
Type:                2


Copy Number:        2
Fragment Size (KB): 6188832
Expires:            1/5/2024 3:19:13 AM
MediaID:            XXX244
Comp:               BBB
Block:              45
Duplicate:          N


Copy Number:        3
Fragment Size (KB): 6188832
Expires:            1/5/2024 3:19:13 AM
MediaID:            XXX199
Comp:               AA
Block:              334

Copy Number:        4
Fragment Size (KB): 6188832
Expires:            1/5/2024 3:19:13 AM
MediaID:            XXX177

到目前为止我有代码：

Get-Content C:\test.txt | Select-String -Pattern 'Backup-ID: ' ,'Policy: ' ,'Primary Copy: ' ,'Expires:  ' ,'Copy Number: ' , 'Fragment Size ' ,'Expires: ' , 'MediaID:'

这就是我想要的：

hostname01,VM_weekly,23,6188832,1/5/2024 3:19:13 AM,XXX122,3,6188832,1/5/2024 3:19:13 AM,XXX134
hostname02,VM_weekly2,24,1/5/2024 3:19:13 AM,2,6188832,1/5/2024 3:19:13 AM,XXX244,3,6188832,1/5/2024 3:19:13 AM,XXX199,4,6188832,1/5/2024 3:19:13 AM,XXX177

Answer 1

这是我的老派方法：

$line = ''
Get-Content C:\test.txt | 
    Select-String -Pattern 'Backup-ID: ' ,'Policy: ' ,'Primary Copy: ' ,'Expires:  ' ,'Copy Number: ' , 'Fragment Size ' ,'Expires: ' , 'MediaID:' |
        ForEach-Object {
            $aux = $_  -split ':',2            # only 2 substrings
            if ($aux[0] -eq 'Backup-ID') {
                if ( $line -ne '' ) { $line }  # Write-Output (current line)
                $line = $aux[1].Trim()
            } else {
                $line += ',' + $aux[1].Trim()
            }
        }
        $line                                   # Write-Output (last line)

输出：

D:\PShell\SO\54921319.ps1

hostname01，VM每周，23，1/5/2024 3:19:13，2,6188832，1/5/2024 3:19，XXX122,3,6188832,1/5/2024 3:19:13上午，XXX134 hostname02，2，24，1/5/2024 3:19:13 AM,2,6188832,1/5/2024 3:19:13 AM，XXX177 244，3,6188832，1/5/2024 3:19:13 AM，XXX177 199，4,6188832，1/2024 3:19:13，XXX177

编辑：…我需要导出CSV文件…。

$xArr = D:\PShell\SO\54921319.ps1
$xCsv = $xArr |  ConvertFrom-Csv -Header $(1..30|%{"a$_"})
$xcsv | Export-Csv -NoTypeInformation -Path c:\temp\result.csv

当然，你可以计算

• -Header $(1..30|%{"a$_"})的实际上限，而不是估计的30，例如($xArr | % {$_.Split(',').Count}|Measure-Object -Maximum).Maximum，
• 甚至计算一些人类可读的头部(考虑到给定Copy Number中每个Backup-ID的属性的反复出现的名称)

Answer 2

使用更好的模式

 $Pattern = '^Backup-ID|^Policy|^Primary Copy|^Expires|^Copy Number|^Fragment Size|^Expires|^MediaID'

和RegEx来拆分Backup-ID的输出。

(Get-Content .\test.txt|Select-String -Pattern $Pattern|Out-String) -split "(?=Backup-ID)"|ForEach-Object {
    (($_ -split "`r?`n" | %{($_ -split ":\s+",2)[1]}) -join ',').Trim(',')
}

hostname01,VM_weekly,23,1/5/2024 3:19:13 AM,2,6188832,1/5/2024 3:19:13 AM,XXX122,3,6188832,1/5/2024 3:19:13 AM,XXX134
hostname02,VM_weekly2,24,1/5/2024 3:19:13 AM,2,6188832,1/5/2024 3:19:13 AM,XXX244,3,6188832,1/5/2024 3:19:13 AM,XXX199,4,6188832,1/5/2024 3:19:13 AM,XXX177

Answer 3

也许是这个？

& {
    $current = $null
    switch -regex -file 'C:\text.txt' {
        '^(Backup-ID|Policy|Primary Copy|Expires|Copy Number|Fragment Size \(KB\)|Expires|MediaID):\s+(.*)' {
            if ($matches[1] -eq "Backup-ID") {
                if ($current) { $current.ToString() }
                $current = [Text.StringBuilder]::new()
                [void]$current.Append($matches[2])
            }
            else {
                [void]$current.Append(",").Append($matches[2])
            }
        }
    }
    $current.ToString()
}