Scan txt file for multiple strings and save the following lines - powershell

Question

I have a problem that I am trying to solve, however, due to my non existing PowerShell knowledge it is proving to be harder than I hoped. So any help would be appreciated.
The problem can be simplified as:
Find a string in a txtfile
Extract the information on the row after that string
Store the information in a handle
Find a second string in the txtfile and repeat the procedure
Store both strings in a new file or delete everything else in the txt file.
I am then trying to do this for approx 20k files. I would love to have the information under their keyword and comma delimited so that I can import them in other systems.
My files look somewhat like the following
random words
that are unimportant
Keyword
FirstlineofNumbersthatIwanttoExtract
random words again that are unimportant
Secondkeyword
SecondLineOfNumbersThatIWantToExtract
end of the file
All files are however not similar in terms of the row that the lines I want to extract are on. I would the output to be something like
Keyword, SecondKeyword
FirstLineOfNumbersThatIWantToExtract, SecondLineOfNumbersThatIWantToExtract
And done. I got this far
[System.IO.DirectoryInfo]$folder = 'C:\users\xx\Desktop\mappcent3'
foreach ($file in ($folder.EnumerateFiles())) {
if ($file.Extension -eq '.txt') {
$content = Get-Content $file
$FirstRegex = 'KeyWordOne
(.+)$'
$First_output = "\1"
$test = Select-String -Path $file.FullName -Pattern $FirstRegex
}
}

Answer 1

This would do something similar to what you are asking. This requires PowerShell 3.0+
$path = 'C:\users\xx\Desktop\mappcent3'
$firstKeyword = "Keyword"
$secondKeyword = "Secondkeyword"
$resultsPath = "C:\Temp\results.csv"
Get-ChildItem $path -Filter "*.txt" | ForEach-Object{
# Read the file in
$fileContents = Get-Content $_.FullName
# Find the first keyword data
$firstKeywordData = ($fileContents | Select-String -Pattern $firstKeyword -Context 0,1 -SimpleMatch).Context.PostContext[0]
# Find the second keyword data
$secondKeywordData = ($fileContents | Select-String -Pattern $secondKeyword -Context 0,1 -SimpleMatch).Context.PostContext[0]
# Create a new object with details gathered.
[pscustomobject][ordered]#{
File = $_.FullName
FirstKeywordData = $firstKeywordData
SecondKeywordData = $secondKeywordData
}
} | Export-CSV $resultsPath -NoTypeInformation
Select-String is what does most of the magic here. We take advantage of -Context which consumes lines before and after the match. We want the one following so that is why we use 0,1. Wrap that up in a custom object and then we can export it to a CSV file.
Keyword Overlap
Beware that your keywords can overlap and create odd results in your output files. In your sample Keyword matches multiple lines so the result set would reflect that.
If you did just want to write back to the original file you could easily do that as well
"$firstKeywordData,$secondKeywordData" | Set-Content $_.FullName
Or something similar.

Answer 2

The Select-String cmdlet has a -Context parameter that makes it easy to extract lines before or after the line on which there's a match.
You can use Export-Csv to export to the format you require (although with 20K files you may want to write directly to the output files)
foreach($file in Get-ChildItem C:\users\xx\Desktop\mappcent3 |Where {-not $_.PsIsContainer})
{
$FirstKeyword = 'FirstKeyword'
$FirstLine = Select-String -Path $file.FullName -Pattern $FirstKeyword -Context 0,1 |Select -Expand Context -First 1 |Select -Expand PostContext
$SecondKeyword = 'SecondKeyword'
$SecondLine = Select-String -Path $file.FullName -Pattern $SecondKeyword -Context 0,1 |Select -Expand Context -First 1 |Select -Expand PostContext
New-Object psobject -Property #{$FirstKeyword=$FirstLine;$SecondKeyword=$SecondLine} |Export-Csv (Join-Path $file.DirectoryName ($file.BaseName + '_keywords.txt'))
}