从文件行删除

Question

I have a text file in which the info id organized in block of lines. I need to delete the blocks that contain some information.

For example I have the following

dn: CN=Publishers,OU=ABC - Groups 
changetype: add 
cn: Cert 
description:   Members of this group are permitted to ... 
groupType: -2 
objectClass: top 
objectClass: group

dn: CN=Domain Guests,OU=ABC - Groups, 
changetype: add 
cn: Domain Guests 
description: All domain guests 
groupType: -21 
objectClass: top
objectClass: group

dn: CN=Domain Computers,OU=ABC- Groups 
changetype: add 
cn: Domain Computers 
description: All workstations  
groupType: -2 
objectClass:top 
objectClass: group

dn: CN=AS Servers,OU=ABC- Groups 
changetype: add 
cn: AS   Servers
description: Servers in this group... 
groupType: -214
objectClass: top 
objectClass: group

dn: CN=Domain Controllers,OU=ABC - Groups 
changetype: add 
cn: Domain Controllers 
description: All domain controllers in the domain
groupType: -21 
objectClass: top 
objectClass: group

dn: CN=Domain Users,OU=ABC - Groups 
changetype: add 
cn: Domain Users
description: All domain users 
groupType: -21 
objectClass: top
objectClass: group

And I need to remove for example the blocks which have the name contained in another file for example the first line for some blocks are in othe file (eg CN=Domain Computers,OU=ABC- Groups and CN=AS Users,OU=ABC- Groups (those bellow)

My code :

$listsharedCN=Get-content "shared.txt
$exported_groups= Get-Content "groups.txt
$listsharedCN | % {
$var=($_ -split '`n')[0]

if(($exported_groups | % { ($_ -split '`n')[0] }) -match[regex]::Escape($var)) 
 {
    #I found the first line of the block but need to delete this block  "
 }  

Can anyone help me?

Answer 1

dn:\s+CN=(?:Domain\s+Computers|AS\s+Users),OU=ABC\s*-\s+Groups.*?\n

You can try this.Replace by empty string .See demo.

http://regex101.com/r/sU3fA2/10

Answer 2

It's not totally clear what you are trying to do, so I will give you some general guideline. You could turn this file into a CSV formatted file and then read it in with ConvertFrom-Csv. You then have an array of proper objects to work with.

Example (you can put all of this on one line):

Get-Content .\ExportGroups.ldf | % { $_ -replace ' (\w+:)', ';' -replace 'dn: ', ''} | 
  ConvertFrom-Csv -Delimiter ';' -Header DN, ChangeType, CN, Description, GroupType, ObjectClass, ObjectClass2

I'll explain it, because it's a bit cryptic.

• Get all the lines of the file with Get-Content
• Use -replace with regex to create valid CSV formatted lines
• Create PowerShell objects from the CSV with ConvertFrom-CSV

Answer 3

While I am not sure what exactly you are doing with this data it would be tonnes easier to manipulate it if it was an Object....... LETS DO THAT! From comments: I feel like I over complicated the issue after I found out what you were doing with the data after the fact. While this should work I would suggest that if you ever have a question to put some code with input data ( if required ) and desired out.

$rawFile = Get-Content -Raw -Path E:\temp\data.txt
$excludesFile = Get-Content -Path E:\temp\shared.txt

$ldifdes = $rawFile -split '\s+(?=dn:)' | ForEach-Object{
    # Empty hashtable that we will use to build a single custom object
    $props = @{}

    $entityObject = $_ -split "`r`n" 

    # Parse out the object class. Custom object cannot have more than one key
    # so we convert to Array instead.
    $entityObject | Where-Object{$_ -notmatch 'objectclass'} | ForEach-Object{
        # Populate the object that are not objectclass into $props
        $parameter = $_ -split ":"
        $props.($parameter[0]) = ($parameter[1]).Trim()
    }

    # Address the objectclass now.
    $props.ObjectClass = @()
    $entityObject | Where-Object{$_ -match 'objectclass'} | ForEach-Object{
        $props.ObjectClass += (($_ -split ":")[1]).Trim()
    } 

    New-Object -TypeName pscustomobject -Property $props
}

# Filter out the entries we do not need.
$regex = "($($excludesFile -join "|"))"
$ldifdes | Where-Object{$_.DN -notmatch $regex} | ForEach-Object{
    Write-Output "dn: $($_.DN)"
    Write-Output "changetype: $($_.changetype)"
    Write-Output "cn: $($_.cn)"
    Write-Output "description: $($_.description)"
    Write-Output "groupType: $($_.groupType)"
    $_.ObjectClass | ForEach-Object{
        Write-Output "ObjectClass: $_"
    }
    # Add a blank
    Write-Output ""
} | Out-File -Encoding ascii -FilePath "E:\temp\output.txt"

If you have at least powershell 3.0 this code will work. Not a big deal to downgrade if need be. Without going into to much detail:

1. Read the files raw contents
2. Split those contents into there variable groups
3. For each one of those group split again into a string array
4. For each string in the string array split by colon to get the key and value
5. Exception being ObjectClass which we convert into an array since a hashtable need to have unique keys
6. Build a regex of items we want to exclude and parse them out with a Where-Object
7. For simplicity sake ( since i feel i made this worse ) use a series of Write-Ouput 's to display the data for ldifde
8. Pipe into Out-File

Less Complicated

$rawFile = Get-Content -Raw -Path E:\temp\data.txt
$excludesFile = Get-Content -Path E:\temp\shared.txt
$regex = "($($excludesFile -join "|"))"

$results = $rawFile | select-string -pattern '(?smi)(dn:).*?(?=dn:|\Z)' -AllMatches | Foreach {$_.Matches} | ForEach-Object {$_.Value.Trim()}

$results | Where-Object{ $_ -notmatch $regex } | ForEach-Object{$_ + [Environment]::NewLine} | 
        Out-File -Encoding ascii -FilePath "E:\temp\output.txt"

Parse the strings into block groups. Build $regex string just like before and use that as an exclusion in Where-Object . Then just output to file using Out-File