如何在PowerShell中删除CSV指定行且无需指定矩阵长度？

Hey there! I get that dealing with huge CSV files and mismatched fields can be frustrating. Let's break down how to delete that problematic line (or any specific line) in PowerShell without having to define a fixed range of lines to keep.

Option 1: Simple Line Filtering with ForEach-Object #

This method is straightforward and works well for most cases. We'll track the line number and skip only the one you need to remove:

# Define the line number you want to delete (matches the error line from R)
$lineToDelete = 458945

# Read the CSV, skip the target line, and write to a new file
Get-Content "C:/a/b/c/file.csv" | 
    ForEach-Object -Begin { $currentLine = 0 } -Process {
        $currentLine++
        if ($currentLine -ne $lineToDelete) { $_ }
    } | 
    Set-Content "output.csv"

Note: Double-check that the line number matches between R's error message and PowerShell. R's fread counts the header as line 1 if your CSV has one, which aligns with how this script counts lines, so you should be safe using the line number from the warning.

Option 2: Memory-Efficient Processing for Large Files #

For 3.5GB files, loading the entire content into memory (like Get-Content does) might strain your system. Instead, use .NET's stream readers/writers to process the file line-by-line with minimal memory footprint:

$inputFilePath = "C:/a/b/c/file.csv"
$outputFilePath = "output.csv"
$lineToDelete = 458945

# Initialize stream readers/writers
$reader = [System.IO.StreamReader]::new($inputFilePath)
$writer = [System.IO.StreamWriter]::new($outputFilePath)

try {
    $currentLine = 0
    while (-not $reader.EndOfStream) {
        $line = $reader.ReadLine()
        $currentLine++
        # Only write the line if it's not the one we want to delete
        if ($currentLine -ne $lineToDelete) {
            $writer.WriteLine($line)
        }
    }
}
finally {
    # Make sure we clean up resources even if something goes wrong
    $reader.Close()
    $writer.Close()
}

This method is much more efficient for large files since it never loads the entire CSV into memory—just processes each line as it reads it.

Quick Side Note About Your R Issue #

Just in case you haven't tried it: you could also tell fread to skip that problematic line directly using the skip parameter, or use badlines() to identify and exclude bad rows. But since you asked specifically about PowerShell, we focused on that solution first!

内容的提问来源于stack exchange，提问作者Earl Mascetti