PowerShell: How Pipe Works

Example of Pipe

You can send one command's output to another command's input. This is called piping. you write it as: command1 | command2 | command3.

# list current dir, sort it
dir | sort

# list, sort, show first 5 elements
dir | sort | select -first 5

# move jpg files to another dir
dir -file *jpg | move ~/Pictures/

# filter out bot access from web log, save to new file
Get-Content weblog | Select-String -notmatch "Googlebot" | Out-File path -width 111222333 -NoNewLine weblog2 

How Pipe Works

• The output of PowerShell commands are dotnet objects. They display on screen as formatted text representation. Unlike unix shells, it is not passing text streams.
• Each dotnet object has a type, and members. Member means property or method. [see Object Type, Properties, Methods]
• Output of a command may be a collection object. (collection is a type, and is like a list, each element is a object.) e.g. output of dir.
• Objects in collection may be different types. e.g. output of dir may be a single System.IO.FileInfo or System.IO.DirectoryInfo, or a collection of them.
• Pipeline output is passed as arguments into one of the paramater of the receiving command. The command's doc (help cmd) specifies which paramaters (if any) can accept argument from pipeline. When more than one paramater can accept pipeline input, PowerShell makes a choice automatically based on the object type. User cannot chose which parameter to use for pipeline input. (unlike unix, the pipeline does not go thru stdout and stdin.)
• When output is a collection, it is piped into receiving command element at a time, in order. (Note: it is not passing the whole collection as one single value) For example, dir | select -First 5.
• If a command receives collection type argument from its paramater and not via pipeline, it is received as one single object. For example, select -InputObject (dir) -First 5 no work.

See also: Object Type, Properties, Methods