Hello,

I would like to know what is the best method for getting a number of
specified string occurrences in a binary file. After some pondering I've made
something like this:

$sum = 0; (Select-String -AllMatches foo .\foobarbaz.txt) |% { $sum +=
$_.Matches.count }; $sum

Where foobarbaz.txt is:
foo bar foo
baz
foo foo bar baz
bar foo bar foo baz
foo

But probably you can recommend some other method. It should be optimal for
large binary files. Or a single command in it's simplest form.

Thanks in advance

You're on the right path, Select-String is the Cmdlet for the task. Wrapping the Select-String statement in parenthesis collects all objects before passing them on to ForEach-Object, this affects performance, specially when processing large files/collections. Optionally, you can initialize $Sum in FoEach-Object's -Begin ScriptBlock and output it in ForEach-Object's -End ScriptBlock.

Using a couple of ,Net methods -IO.File's ReadAllText and Regex's Matches- can speed up the process dramatically but your RAM will suffer, but you can cleanup with another .Net method, GC's Collect. You could write up a function to make it simpler.

# both one-liners, look out for word wrapping...
# pure PowerShell
Select-String -AllMatches foo .\foobarbaz.txt | % {$sum = 0} {$sum += $_.Matches.count} {$sum}

# .Net through PowerShell, note the fullpath of the file
[regex]::Matches([IO.File]::ReadAllText('c:\somedir\foobarbaz.txt'), 'foo', 'IgnoreCase').count; [gc]::Collect()

IO.File's ReadAllText Method is fast. More about it here:
http://msdn.microsoft.com/en-us/libr...adalltext.aspx

Regex's Matches Method is casesensitive, use the IgnoreCase option. More here:
http://msdn.microsoft.com/en-us/libr...x.matches.aspx

GC's Collect:
http://msdn.microsoft.com/en-us/libr...c.collect.aspx

--
Kiron

# ...or
([IO.File]::ReadAllText('c:\somedir\foobarbaz.txt') | Select-String foo -all).matches.count

--
Kiron

Kiron,

Many thanks for your suggestions! I think I will stick with your version of
pure PowerShell one-liner.

Dainis

BTW -AllMatches switch is new to PowerShell v2.0. Seems that v1.0 users
should use the first .NET solution:

[regex]::Matches([IO.File]::ReadAllText('c:\somedir\foobarbaz.txt'), 'foo',
'IgnoreCase').count; [gc]::Collect()