使用PowerShell，RegEx搜索在文本文件中查找和验证路径字符串

Question

First time posting here, and I'll try to be clear and detailed, but be gentle if I missed an existing answer when I searched these boards.

First, the issues:

1. How to exclude a RegEx response that contains a specific keyword ("fastcopy")
2. How to include path results that do not end in a file name/wildcard

I am working with a set of text files that are very similar to batch files. They are plain text, and contain header lines, lines containing paths to files on a server, and comment lines. Commented lines begin with a semicolon (;), so that is simple enough to rule out. The paths should all start with a variable %INSTDIR%, but they may or may not have quotes surrounding the path, and they may or may not have execution options following the path. One last note... the company uses FastCopy.exe to dump files/folders down from the network, and in such a line I would like to return the folder/file being copied instead of the path containing fastcopy.exe.

Here is a sample (kind of large to show potential issues):

[Installing .NET 3.5 Hotfix KB943326 for App1]
; *** Added NET 3.5 SP1 hotfix KB943326: resolves App1 hidden menus force laptop re-booting
1 = %INSTDIR%\ToolShare$\Sample_Toolbox\applications\.NET_3.5_Hotfix_KB943326\WindowsXP-KB943326-x86-ENU.exe /quiet /norestart

[Installing Agent 5.3.1]
1 = %INSTDIR%\ToolShare$\Sample_Toolbox\applications\AGenT_531_2.0\w7wxp_ze_20\install.exe

[Installing APR Manager 2.1]
1 = %INSTDIR%\ToolShare$\Sample_Toolbox\applications\APRManager_21_Updated_2.0\wviwxp_ze_20\install.exe

[Installing Scope Simulator]
1 = MD "C:\Temp\scope_simulator_10"
2 =  start /wait /high %INSTDIR%\ToolShare$\Site_Toolbox\Custom_Scripts\Source\fastcopy.exe /auto_close /no_confirm_del /no_confirm_stop /log=FALSE /open_window /force_start /force_close /stream=FALSE /cmd=diff "%INSTDIR%\ToolShare$\Sample_Toolbox\applications\scope_simulator_10" /to="C:\Temp\scope_simulator_10"
3 = "C:\Temp\scope_simulator_10\w7wxp_ze_10\Install.exe"
4 = RD "C:\temp\scope_simulator_10" /q /s

[Installing Log Analyzer Offline 2.6.1]
1 = %INSTDIR%\ToolShare$\Sample_Toolbox\applications\Log_Analyzer_Offline_261\wxp_ze_10\install.exe

[Installing Data Migration Script]
1 = MD "C:\Temp\Data Migration"
2 = xcopy "%INSTDIR%\ToolShare$\Sample_Toolbox\Support\Data Migration\*.*" "C:\Temp\Data Migration" /y /e
3 = xcopy "%INSTDIR%\ToolShare$\Sample_Toolbox\Support\Data Migration\Data Migration.lnk" C:\DOCUME~1\ALLUSE~1\Desktop\ /Y

I have it set to pull a 'dir \\\\UNCPath\\*.ini' and then loop through that doing a ForEach ($INI in $Results) bit. The line that I have been using inside the loop to try and pull the paths from each line is:

gc $ini|?{!($_ -match "^;") -and ($_ -match "%INST[^`"]*?\\.*(\.\w{3}|\.\*)(?=`"|\s|\Z)")}|%{$TestPath = $Matches[0].replace("%INSTDIR%","\\ServerName1");if(test-path $testpath){write-host "  [OK]    " -foregroundcolor Green -NoNewline}else{write-host "[Missing] " -ForegroundColor red -NoNewline};write-host "$testpath"}

This gets me almost everything I could want. What it doesn't do is get anything that does not end in either a .* or standard 3 character extension (.exe, .cmd, .jar etc). Plus it kicks back the fastcopy path instead of the path that it being attempted to be copied.

What I would like for results:

%INSTDIR%\ToolShare$\Sample_Toolbox\applications\.NET_3.5_Hotfix_KB943326\WindowsXP-KB943326-x86-ENU.exe
%INSTDIR%\ToolShare$\Sample_Toolbox\applications\AGenT_531_2.0\w7wxp_ze_20\install.exe
%INSTDIR%\ToolShare$\Sample_Toolbox\applications\APRManager_21_Updated_2.0\wviwxp_ze_20\install.exe
%INSTDIR%\ToolShare$\Sample_Toolbox\applications\scope_simulator_10
%INSTDIR%\ToolShare$\Sample_Toolbox\applications\Log_Analyzer_Offline_261\wxp_ze_10\install.exe
%INSTDIR%\ToolShare$\Sample_Toolbox\Support\Data Migration\*.*
%INSTDIR%\ToolShare$\Sample_Toolbox\Support\Data Migration\Data Migration.lnk

I do not get the second result (instead I get the FastCopy path, but even if I strip Fastcopy from the line and only have the desired path it won't return it). Any suggestions are welcome.

Answer 1

The following script should work just fine.

$paths = Get-Content $ini | Foreach {
    if ($_ -match "^(?=[^;]).*?(?<delimiter>[""' ])(?<path>%INSTDIR%(?!.*?fastcopy.exe).*?)(?:\1|$)")
    {
        Write-Output $Matches["path"]
    }
}

The $paths variable will now contain all the paths requested. Observe that if any string contains the "fastcopy.exe" literal string anywhere in the path it will not be found by this regular expression.

An attempt to explaining the regular expression:

^ - match the start of the line
(?=[^;]) - positive lookahead verifying that the line does not start with a semicolon
.*? - any character, as few as possible (to remove all characters before the path we want to match)
(?<delimiter>["' ]) - named group verifying whether the path is surrounded by space, a quotation character or a apostrophe.
(?<path> - start a named capturing group for capturing the "path"
    %INSTDIR% - matches the literal string '%INSTDIR%'
    (?!.*?fastcopy.exe) - negative lookahead verifying that the part of the line we're trying to match (which has started with %INSTDIR%) doesn't contain the word fastcopy.exe anywhere later in the string (the second time the %INSTDIR% occurs on the fastcopy line, the rest of the line does not contain the fastcopy.exe literal string).
    .*? - matches any character, as few as possible, to make sure that we stop as soon as we find a matching delimiter character below
) - ends the named capturing group "path"
(?:\1|$) - matches (in a non-capturing group) the character found by the delimiter group above (to match a quotation character, apostrophe or space, depending on what character was immediately before the %INSTDIR% literal string), or the end of the line.

If anything is unclear, please add a comment below asking for clarifications.