Group extensions by size

I have a sample that does extensions by count for my F: drive, but how
do I do this by length instead of count?

get-childitem -r c:\ | group extension | sort-object count -descending
| select -first 10

Thank you in advance

"Ant" <antgoodlife@gmail.com> wrote in message
news:1176136308.451747.197050@y80g2000hsf.googlegroups.com...
>I have a sample that does extensions by count for my F: drive, but how
> do I do this by length instead of count?
>
> get-childitem -r c:\ | group extension | sort-object count -descending
> | select -first 10

Probably not optimised for a complete drive scan, but basically equivalent
to your sample:

dir -r c:\ | group extension | select @{n="Size";e={($_.group | measure
length -sum).sum}},name | sort size -desc | select -first 10

Jacques

Thank you, while I am sure this is close, it's throwing this error :
PS F:\> dir -r F:\ | group extension | select @{n="Size";e={($_.group
| measure length -sum).sum}},name | sort size -desc | select -first 10
Select-Object : The term 'measure' is not recognized as a cmdlet,
function, operable program, or script file. Verify the term and try
again.
At line:1 char:38
+ dir -r F:\ | group extension | select <<<< @{n="Size";e={($_.group
| measure length -sum).sum}},name | sort size -desc | select -first 10
Select-Object : The term 'measure' is not recognized as a cmdlet,
function, operable program, or script file. Verify the term and try
again.
At line:1 char:38

over and over....

Also, could you please explain this section so I learn a little
without just getting the fish?
select @{n="Size";e={($_.group | measure length -sum).sum}},name

Thank you again...


Thanks again...


On Apr 9, 1:11 pm, "Jacques Barathon [MS]"
<jbara...@online.microsoft.com> wrote:
> "Ant" <antgoodl...@gmail.com> wrote in message
>
> news:1176136308.451747.197050@y80g2000hsf.googlegroups.com...
>
> >I have a sample that does extensions by count for my F: drive, but how
> > do I do this by length instead of count?
>
> > get-childitem -r c:\ | group extension | sort-object count -descending
> > | select -first 10
>
> Probably not optimised for a complete drive scan, but basically equivalent
> to your sample:
>
> dir -r c:\ | group extension | select @{n="Size";e={($_.group | measure length -sum).sum}},name | sort size -desc | select -rst 10
>
> Jacques

> Thank you, while I am sure this is close, it's throwing this error :
> PS F:\> dir -r F:\ | group extension | select @{n="Size";e={($_.group
> | measure length -sum).sum}},name | sort size -desc | select -first 10
> Select-Object : The term 'measure' is not recognized as a cmdlet,
> function, operable program, or script file. Verify the term and try
> again.
> At line:1 char:38

That's odd, it works here okay on Powershell V1.0. Try expanding
'measure' to 'measure-object'

> Also, could you please explain this section so I learn a little
> without just getting the fish?
> select @{n="Size";e={($_.group | measure length -sum).sum}},name
>
> Thank you again...
>

I can't quite understand the hash-literal in the mid-section either.
I see that n= can be expanded to name=, but any other token generates
an invalid key error, same goes for e=, that must be a short-cut for
something too, enum maybe?

Looks like it's a way of populating the hashtable in-line, but can't
find the relevant documentation.

Thanks,

Duncan.

Yes, measure-object was the trick.. (And I am using powershell version
1.0 too...) perhaps you have an alias... regardless I'm all set.
Thanks again for your help, now onto deleting linux gzips from my home
drive :-)

PS F:\> dir -r F:\ | group extension | select @{n="Size";e={($_.group
| measure-object length -sum).sum}},name | sort size -desc | select -
first 10


Size Name

---- ----

348529270 .gz

98309149 .exe

47478784 .nsf

47379546 .pdf

36364346 .jar

30728192 .ntf

29310093 .doc

16187033 .dll

10936641 .rtf

10219727 .txt


On Apr 9, 3:12 pm, "Duncan Smith" <DSmith1...@googlemail.com> wrote:
> > Thank you, while I am sure this is close, it's throwing this error :
> > PS F:\> dir -r F:\ | group extension | select @{n="Size";e={($_.group
> > | measure length -sum).sum}},name | sort size -desc | select -first 10
> > Select-Object : The term 'measure' is not recognized as a cmdlet,
> > function, operable program, or script file. Verify the term and try
> > again.
> > At line:1 char:38
>
> That's odd, it works here okay on Powershell V1.0. Try expanding
> 'measure' to 'measure-object'
>
> > Also, could you please explain this section so I learn a little
> > without just getting the fish?
> > select @{n="Size";e={($_.group | measure length -sum).sum}},name
>
> > Thank you again...
>
> I can't quite understand the hash-literal in the mid-section either.
> I see that n= can be expanded to name=, but any other token generates
> an invalid key error, same goes for e=, that must be a short-cut for
> something too, enum maybe?
>
> Looks like it's a way of populating the hashtable in-line, but can't
> find the relevant documentation.
>
> Thanks,
>
> Duncan.

"Ant" <antgoodlife@gmail.com> wrote in message
news:1176146219.534107.182510@p77g2000hsh.googlegroups.com...
> Yes, measure-object was the trick.. (And I am using powershell version
> 1.0 too...) perhaps you have an alias... regardless I'm all set.
> Thanks again for your help, now onto deleting linux gzips from my home
> drive :-)

Yes, sorry about that. I have a script block in my profile which aliases all
cmdlets having object as their name to their verb only (new-object -> new,
measure-object -> measure, compare-object -> compare, etc).

Re: the in-line hash, I can't tell you where it is documented but I am quite
sure it is somewhere. Others in the PowerShell team may be able to guide you
here. I am aware of those tokens:

* Expression - abbreviated to e.
Contains the code you want to execute, $_ being the current object. This
token is common to all cmdlets that can use such hash. You can only the
expression and nothing else, and then skip the "e=" part of the hash (eg.
select {($_.group | measure-object length -sum).sum}). An example with
another cmdlet than select:
dir | group {"{0}{1}" -f $_.name[0],$_.extension[1]}

* Name - abbreviated to n.
Gives a name to the property you are returning from your expression. I have
used it with select, not sure if it can apply to any other cmdlet.

* Label - abbreviated to l.
Replaces Name in formatting cmdlets.

* Align - abbreviated to a.
Decides of the alignment of text in formatting cmdlets. Can be "left",
"right" or "center". Another example (completely useless, as all great
examples are):
dir | ft @{l="Initials";a="right";e={"{0}{1}" -f
$_.name[0],$_.extension[1]}} -a

There might be other formatting tokens, but I very rarely need them.

Hope that helps,
Jacques