Powershell外文件强制行尾字符

Question

I discovered that I could force a Unicode file to ASCII using the script below, which is really great. I assume it's based on my environment or Windows default, but it's adding a CR and LF at the end of each line. Is there a way to force just a LF character rather than both without loading the entire file into memory? I have seen some solutions that load the entire file into memory and basically do a string replace, which won't work because some of my files are multiple GB.

Thanks!

get-content -encoding utf8 $inputFile | Out-file -force -encoding ASCII $outputFile

Answer 1

I suggest you use .NET System.File.IO classes from within your script. In particular the System.File.IO.StreamWriter class has a property, NewLine which you can set to whatever characters you want the line terminator characters to be. (Although to be readable by StreamReader the line terminator chars must be \\n or \\r\\n (in C/C++ notation because of conflict with SO and PS on backtick)).

Secondary benefit of using IO.StreamWriter, according to this blog is much better perf.

Basic code flow is something like this (not tested):

# Note that IO.StreamWriter will use process's current working directory,
#  not PS's. So safer to specify full paths
$inStream =  [System.IO.StreamReader] "c:\temp\orig.txt"
$outStream = new-object System.IO.StreamWriter "c:\temp\copy.txt",  
                                               [text.encoding]::ASCII
$outStream.NewLine = '`n'
while (-not $inStream.endofstream) {
  $outStream.WriteLine( $instream.Readline())
}
$inStream.close()
$outStream.close()

This script should have constant memory requirements, but hard to know what .NET might do under the covers.

Answer 2

Can't comment, and apparently the edit isn't enough characters, but it's worth pointing out the string literal usage in

$outStream.Newline = '`n'

backfires, because it passes

`n

instead of the newline character itself to $outStream . Should be:

$outStream.Newline = "`n"