Two files: keep lines with identical first n characters only - powershell

Question

There are 2 text files in the CWD, a.txt, b.txt. From a.txt, I would like to delete all lines whose first 5 characters are NOT present in b.txt as any lines' first 5 characters. (Or, stating otherwise, keep only those lines in a.txt, whose first 5 characters is present in b.txt as any lines' first 5 characters.) Content after the 5th character to the end of the line is irrelevant.
For example: a.txt
abcde000dsdsddsdsdsdsdsd
0123456xxx
kkk
xyzxyzxyzfeeeee
kkkkkkkkkkk
and b.txt:
012345aabbcc
kkkkkkkhhkkvv
nnnnnnn5777nnnn77567
Intended result (lines in a.txt whose 1-5 character is present in b.txt):
0123456xxx
kkkkkkkkkkk
When I am running the code, it gives me an empty results.txt, but no error messages. What I am missing?
$pattern = "^[5]"
$set1 = Get-Content -Path a.txt
$results = New-Object -TypeName System.Text.StringBuilder
Get-Content -Path b.txt | foreach {
if ($_ -match $pattern) {
[void]$results.AppendLine($_)
}
}
$results.ToString() | Out-File -FilePath .\results.txt -Encoding ascii

Answer 1

Your code doesn't work because your pattern doesn't match anything. The regular expression ^[5] means "the character '5' at the beginning of the string" (the square brackets define a character class), not "5 characters at the beginning of the string". The latter would be ^.{5}. Also, you never match the content of a.txt against the content of b.txt.
There are several ways to do what you want:
Extract the first 5 characters from each line of b.txt. to an array and compare the lines of a.txt against that array. Esperento57's answer sort of uses this approach, but in a way that requires PowerShell v3 or newer. A variant that'll work on all PowerShell versions could look like this:
$pattern = '^(.{5}).*'
$ref = (Get-Content 'b.txt') -match $pattern -replace $pattern, '$1' |
Get-Unique
Get-Content 'a.txt' | Where-Object {
$ref -contains ($_ -replace $pattern, '$1')
} | Set-Content 'results.txt'
Since lookups in arrays are comparatively slow and don't scale well (they get significantly slower with increasing number of elements in the array) you could also put the reference values in a hashtable so you can do index lookups (which are significantly faster):
$pattern = '^(.{5}).*'
$ref = #{}
(Get-Content 'b.txt') -match $pattern -replace $pattern, '$1' |
ForEach-Object { $ref[$_] = $true }
Get-Content 'a.txt' | Where-Object {
$ref.ContainsKey(($_ -replace $pattern, '$1'))
} | Set-Content 'results.txt'
Another alternative would be to build a second regular expression from the substrings extracted from b.txt and compare the content of a.txt against that expression:
$pattern = '^(.{5}).*'
$list = (Get-Content 'b.txt') -match $pattern -replace $pattern, '$1' |
Get-Unique |
ForEach-Object { [regex]::Escape($_) }
$ref = '^({0})' -f ($list -join '|')
(Get-Content 'a.txt') -match $ref | Set-Content 'results.txt'
Note that each of these approaches will ignore lines shorter than 5 characters.

Answer 2

try Something like this:
$listB=get-content "c:\temp\b.txt" | where {$_.Length -gt 4} | select #{N="First5";E={$_.Substring(0, 5)}}
get-content "c:\temp\a.txt" | where {$_.Length -gt 4 -and $_.Substring(0, 5) -in $listB.First5}

Answer 3

If performance is a concern, consider to use the hashtable(s) as index:
$Pattern = '^(.{5}).*'
$a = #{}; $b = #{}
Get-Content -Path a.txt | Where {$_ -Match $Pattern} | ForEach {$a[$Matches[1]] = #($a[$Matches[1]] + $_)}
Get-Content -Path b.txt | Where {$_ -Match $Pattern} | ForEach {$b[$Matches[1]] = #($b[$Matches[1]] + $_)}
$a.Keys | Where {$b.Keys -Contains $_} | ForEach {$a.$_} | Set-Content results.txt