Pipeline (Unix)

A pipeline of three programs run on a text terminal

In Unix-like computer operating systems, a pipeline is the original software pipeline: a set of processes chained by their standard streams, so that the output of each process (stdout) feeds directly as input (stdin) of the next one. Each connection is implemented by an anonymous pipe. Filter programs are often used in this configuration. The concept was invented by Douglas McIlroy for Unix shells and it was named by analogy to a physical pipeline.

Example

Below is an example of a pipeline that implements a kind of spell checker for the web resource indicated by a URL. An explanation of what it does follows. (Some machines have /usr/share/dict/words instead.)

curl "http://en.wikipedia.org/wiki/Pipeline_(Unix)" | \
sed 's/[^a-zA-Z ]/ /g' | \
tr 'A-Z ' 'a-z\n' | \
grep '[a-z]' | \
sort -u | \
comm -23 - /usr/dict/words

First, curl obtains the HTML contents of a web page (could use wget on some systems).
Second, sed this expression removes all characters which are not spaces or letters from the web page's content, replacing them with spaces.
Third, tr changes all of the uppercase letters into lowercase and converts the spaces in the lines of text to newlines (each 'word' is now on a separate line).
Fourth, grep includes only lines that contain at least one lowercase alphabetical character (removing any blank lines).
Fifth, sort sorts the list of 'words' into alphabetical order, and the -u switch removes duplicates.
Finally, comm finds lines in common between two files, -23 suppresses lines unique to the second file, and those which are common to both, leaving only those which are found only in the first file named. The - in place of a filename causes comm to use its standard input (from the pipe line in this case). This results in a list of "words" (lines) which are not found in /usr/dict/words.
The special character "|" tells the operating system to pipe the output from the previous command in the line into the next command in the line. That is, the output of the curl command is given as the input of the sed command.
The character "\" is used to place all five lines into a single command line.

Pipelines in command line interfaces

Most Unix shells have a special syntax construct for the creation of pipelines. Typically, one simply writes the filter commands in sequence, separated by the ASCII vertical bar character "|" (which, for this reason, is often called "pipe character" by Unix users). The shell starts the processes and arranges for the necessary connections between their standard streams (including some amount of buffer storage).

Error stream

By default, the standard error streams ("stderr") of the processes in a pipeline are not passed on through the pipe; instead, they are merged and directed to the console. However, many shells have additional syntax for changing this behaviour. In the csh shell, for instance, using "|&" instead of "| " signifies that the standard error stream too should be merged with the standard output and fed to the next process. The Bourne Shell can also merge standard error, using 2>&1, as well as redirect it to a different file.

Pipemill

In the most commonly used simple pipelines the shell connects a series of sub-processes via pipes, and executes external commands within each sub-process. Thus the shell itself is doing no direct processing of the data flowing through the pipeline.

However, it's possible for the shell to perform processing directly. This construct generally looks something like:

command | while read var1 var2 ...; do
   # process each line, using variables as parsed into $var1, $var2, etc
   done

... which is referred to as a "pipemill" (since the while is "milling" over the results from the initial command.)

Example:

find / /usr /var -mount -user foo -printf "%m %p\n" | while read mode filename; do
    chown $NEWOWNER "$filename"
    chmod $MODE "$filename"
    done

(This example will traverse file directory trees changing the ownership of all files while preserving all permissions, including those that are often stripped off by many versions of the chown command).

There are a number of variations of the pipemill construct including:

ps lax | { read x; while read x owner pid parent x x x x x stat x; do
   [ "$owner"="foo" -a "$stat"="Z" ] && kill "$parent"
   done
   }

(This example kills the parent processes for zombies owned/created by the user "foo").

Here the while loop is enclosed in a command group (the braces); and preceded by a read command which effectively "throws away" the first line from the ps command. (Of course, in this particular example it would be harmless to process the header line, as it wouldn't match the "$owner"= test). Note that the other references to the "x" variable are simply being used as placeholders for "throwing away" irrelevant fields from each line.

The defining characteristics of a "pipemill" are: some command or series of commands feeds data into a pipe from which a shell while loop reads and processes it.

Creating pipelines programmatically

Pipelines can be created under program control. The pipe() system call asks the operating system to construct a new anonymous pipe object. This results in two new, opened file descriptors in the process: the read-only end of the pipe, and the write-only end. The pipe ends appear to be normal, anonymous file descriptors, except that they have no ability to seek.

To avoid deadlock and exploit parallelism, the process with one or more new pipes will then, generally, call fork() to create new processes. Each process will then close the end(s) of the pipe that it will not be using before producing or consuming any data. Alternatively, a process might create a new thread and use the pipe to communicate between them.

Named pipes may also be created using mkfifo() or mknod() and then presented as the input or output file to programs as they are invoked. They allow multi-path pipes to be created, and are especially effective when combined with standard error redirection, or with tee.

Implementation

In most Unix-like systems, all processes of a pipeline are started at the same time, with their streams appropriately connected, and managed by the scheduler together with all other processes running on the machine. An important aspect of this, setting Unix pipes apart from other pipe implementations, is the concept of buffering: a sending program may produce 5000 bytes per second, and a receiving program may only be able to accept 100 bytes per second, but no data are lost. Instead, the output of the sending program is held in a buffer, or queue. When the receiving program is ready to read data, the operating system sends it data from the buffer, then removes that data from the buffer. If the buffer fills up, the sending program is suspended (blocked) until the receiving program has had a chance to read some data and make room in the buffer.

Network pipes

Tools like netcat and socat can connect pipes to TCP/IP sockets, following the Unix philosophy of "everything is a file".

History

The pipeline concept and the vertical-bar notation was invented by Douglas McIlroy, one of the authors of the early command shells, after he noticed that much of the time they were processing the output of one program as the input to another. The idea was eventually ported to other operating systems, such as DOS, OS/2, Windows NT, and BeOS, often with the same notation.

The robot in the icon for Apple's Automator, which also uses a pipeline concept to chain repetitive commands together, holds a pipe as recognition of the application's Unix heritage.^[1]

Other operating systems

This feature of Unix was borrowed by other operating systems, such as Taos and MS-DOS, and eventually became the pipes and filters design pattern of software engineering.

External links

Pipes: A Brief Introduction by The Linux Information Project (LINFO)
[1] Unix Pipes -- powerful and elegant programming paradigm (Softpanorama)
Ad Hoc Data Analysis From The Unix Command Line at Wikibooks shows how to use pipelines composed of simple filters to do complex data analysis.
stdio buffering
Use And Abuse Of Pipes With Audio Data gives an introduction to using and abusing pipes with netcat, nettee and fifos to play audio across a network.
hitting the pipe A program that forks two processes that communicate with each other using pipes.

References

^ Cite error: The named reference automator was invoked but never defined (see the help page).

Sal Soghoian on MacBreak Episode 5 "Enter the Automatrix"

[automator-1] Cite error: The named reference automator was invoked but never defined (see the help page).

[1]