Get-Content and Measure-Object are fine for small files, but both are super inefficient with memory. I had real problems with large files.
When counting rows in a 1GB file using either method, Powershell gobbled up all available memory on the server (8GB), then started paging to disk. I left it over an hour, but it was still paging to disk so I killed it.
The best method I found for large files is to use IO.StreamReader to load the file from disk and count each row using a variable. This keeps memory usage down to a very reasonable 25MB and is much, much quicker, taking around 30 seconds to count rows in a 1GB file or a couple of minutes for a 6GB file. It never eats up unreasonable amounts of RAM, no matter how large your file is:
[int]$LinesInFile = 0
$reader = New-Object IO.StreamReader 'c:filename.csv'
while($reader.ReadLine() -ne $null){ $LinesInFile++ }
The above snippet can be inserted wherever you would use get-content or measure-object, simply refer to the $LinesInFile variable to get the row count of the file.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…