I want to read a file line by line in PowerShell. Specifically, I want to loop through the file, store each line in a variable in the loop, and do some processing on the line.
I know the Bash equivalent:
while read line do
if [[ $line =~ $regex ]]; then
# work here
fi
done < file.txt
Not much documentation on PowerShell loops.
Not much documentation on PowerShell loops.
Documentation on loops in PowerShell is plentiful, and you might want to check out the following help topics: about_For, about_ForEach, about_Do, about_While.
foreach($line in Get-Content .\file.txt) {
if($line -match $regex){
# Work here
}
}
Another idiomatic PowerShell solution to your problem is to pipe the lines of the text file to the ForEach-Object cmdlet:
Get-Content .\file.txt | ForEach-Object {
if($_ -match $regex){
# Work here
}
}
Instead of regex matching inside the loop, you could pipe the lines through Where-Object to filter just those you're interested in:
Get-Content .\file.txt | Where-Object {$_ -match $regex} | ForEach-Object {
# Work here
}
gc 'file.txt' | ?{ $_ -match $regex } | %{ <#stuff#> } –
Blackfish Get-Content ... |ForEach-Object { ... } has very different performance characteristics than, say (Get-Content ...) |ForEach-Object { ... }, at least for very large files. –
Lou Get-Content .\file.txt | ForEach-Object { ... } :) –
Lou (Get-Content ...) (ie. blocking the Get-Content call) would have very different performance characteristics. The very first example is foreach($line in Get-Content ...){ ... }, and similarly to the second example, foreach will start consuming output from Get-Content as soon as it's available. Feel free to post your own answer if you feel that mine is incomplete –
Lou $date2 value you obtained 25 seconds earlier. Try adding a break statement after $date2 = get-date and you'll find it's near instant - proving that Get-Content does not complete reading the file before loop iteration starts –
Lou Get-Content has bad performance; it tries to read the file into memory all at once.
C# (.NET) file reader reads each line one by one
Best Performace
foreach($line in [System.IO.File]::ReadLines("C:\path\to\file.txt"))
{
$line
}
Or slightly less performant
[System.IO.File]::ReadLines("C:\path\to\file.txt") | ForEach-Object {
$_
}
The foreach statement will likely be slightly faster than ForEach-Object (see comments below for more information).
[System.IO.File]::ReadLines("C:\path\to\file.txt") | ForEach-Object { ... }. The foreach statement will load the entire collection to an object. ForEach-Object uses a pipeline to stream with. Now the foreach statement will likely be slightly faster than the ForEach-Object command, but that's because loading the whole thing to memory usually is faster. Get-Content is still terrible, however. –
Mozell foreach() is an alias of Foreach-Object –
Particiaparticipant foreach is a statement, like if, for, or while. ForEach-Object is a command, like Get-ChildItem. There is also a default alias of foreach for ForEach-Object, but it is only used when there is a pipeline. See the long explanation in Get-Help about_Foreach, or click the link in my previous comment which goes to an entire article by Microsoft's The Scripting Guys about the differences between the statement and the command. –
Mozell Get-Alias foreach => Foreach-Object, but you are right, there are differences –
Particiaparticipant $line to $_ in the loop's script block. –
Mozell Method invocation failed because [System.IO.File] doesn't contain a method named 'ReadLines'. This answer works fine. –
Undone 2 or below, you should upgrade –
Particiaparticipant Get-Content is the devil. –
Particiaparticipant (Get-Content .\BTSManifest.txt | Join-String -Separator ',') . How to perform such a join using these two methods? –
Workroom Reading Large Files Line by Line
Original Comment (1/2021) I was able to read a 4GB log file in about 50 seconds with the following. You may be able to make it faster by loading it as a C# assembly dynamically using PowerShell.
[System.IO.StreamReader]$sr = [System.IO.File]::Open($file, [System.IO.FileMode]::Open)
while (-not $sr.EndOfStream){
$line = $sr.ReadLine()
}
$sr.Close()
Addendum (3/2022) Processing the large file using C# embedded in PowerShell is even faster and has less "gotchas".
$code = @"
using System;
using System.IO;
namespace ProcessLargeFile
{
public class Program
{
static void ProcessLine(string line)
{
return;
}
public static void ProcessLogFile(string path) {
var start_time = DateTime.Now;
StreamReader sr = new StreamReader(File.Open(path, FileMode.Open));
try {
while (!sr.EndOfStream){
string line = sr.ReadLine();
ProcessLine(line);
}
} finally {
sr.Close();
}
var end_time = DateTime.Now;
var run_time = end_time - start_time;
string msg = "Completed in " + run_time.Minutes + ":" + run_time.Seconds + "." + run_time.Milliseconds;
Console.WriteLine(msg);
}
static void Main(string[] args)
{
ProcessLogFile("c:\\users\\tasaif\\fake.log");
Console.ReadLine();
}
}
}
"@
Add-Type -TypeDefinition $code -Language CSharp
PS C:\Users\tasaif> [ProcessLargeFile.Program]::ProcessLogFile("c:\\users\\tasaif\\fake.log")
Completed in 0:17.109
The almighty switch works well here:
'one
two
three' > file
$regex = '^t'
switch -regex -file file {
$regex { "line is $_" }
}
Output:
line is two
line is three
Set-Location 'C:\files'
$files = Get-ChildItem -Name -Include *.txt
foreach($file in $files){
Write-Host("Start Reading file: " + $file)
foreach($line in Get-Content $file){
Write-Host($line)
}
Write-Host("End Reading file: " + $file)
}
© 2022 - 2024 — McMap. All rights reserved.
Get-Contentloads the entire file into memory at once, which will fail or freeze on large files. – Particiaparticipantprocessblock, and spits out another object per line into the pipeline, then that function is the problem. Any problems with loading the full content into memory are not the fault ofGet-Content. – Whensoeverforeach($line in Get-Content .\file.txt)It will load the entire file into memory before it begins iterating. If you don't believe me, go get a 1GB log file and try it. – ParticiaparticipantGet-Content .\file.txt | ForEach-Object -Process {}is pipeline aware, and will not load the entire file into memory. By default Get-Content will pass one line at a time through the pipeline. – Whensoever