Introduction

Command shell is the basic shell in Windows and one of the two preinstalled shells (the other is PowerShell).

The commands available in the Command shell are refered to as Windows commands.

As the PowerShell was designed as an extension of the Command shell, the Windows commands can be executed in PowerShell.

Script files for the Command shell are often called batch files, and have the .bat or .cmd extension.

Comments

Instead of some conventional comment syntax, the Command shell uses the rem command. Example:

rem This is a comment. We print Hello, World! below it.
echo Hello, World!

echo: Displaying text and automatic command printing

The echo command serve two important purposes:

  • displaying text, like in PowerShell and
  • switching on/off the automatic command printing feature of the Command shell.

The automatic command printing basically prints every command that is executed. We can skip the printing by prefixing the command with @. Example:

echo Hello, World! # prints echo Hello, World!\nHello, World!
@echo Hello, World! # prints Hello, World!

However, typically, we want to disable the automatic command printing for the whole script, this can be done by adding call echo with parameter off. To turn it back on, we can call echo with parameter on. So the typical first command in a batch file is:

@echo off

Which disables the automatic command printing and also skips it for the disabeling command itself.

Executable Execution

Executable path resolution

Old reference for Windows NT

Unfortunatelly, there is no up-to-date reference for Command Shell execution logic. However, using the information from the old reference, the execution logic is as follows:

  • If the specified executable is a path, the path is evaluated. If the path is valid, the executable is executed. Otherwise, error is raised.
  • If the specified executable is a name, the name is searched in the locations below (in order):
    1. Windows Commands
    2. Current directory
    3. directories listed in the PATH environment variable, in the order they are listed

When we mention the executable name, we mean the name of the executable file with the extension. Executables names can be also specified without the extension, but only if the extension is listed in the PATHEXT environment variable.

Variables

Variables are set using the set command. Example:

set myVar=Hello

To use a variable, we use its name surrounded by %. Example:

echo %myVar%
rem prints Hello

To display all variables, use the set command without any arguments.

Note that unlike in PowerShell, there is no distinction between local and environment variables. For example, to set the value of the PATH environment variable, we can use the set command:

set PATH=C:\Program Files\Java\jdk1.8.0_181\bin;%PATH%

The set command and its important nuances

The variable setting has several surprising nuances with important consequences:

  • all whitespaces are interpreted and not discarded: ```batch set myVar=Hello rem variable 'myVar' is set to 'Hello'

    set myVar= Hello rem variable 'myVar' is set to ' Hello'

    set myVar =Hello rem variable 'myVar ' is set to 'Hello' ```

  • The quoting for set use a different syntax than expected:

    ```batch set myVar="Hello, World!" rem myVar is set to '"Hello, World!"'

    set "myVar=Hello, World!" rem myVar is set to 'Hello, World!' - no quoting is needed for spaces in the value:batch set myVar=Hello World! rem myVar is set to 'Hello World!' ```

String Manipulation

Unlike in PowerShell or programming languages, there is no concatenation operator in Command Shell. Instead, we just place the strings/variables next to each other. Example:

set myVar=Hello, 
set myVar=%myVar% World!
echo %myVar% # prints "Hello, World!"

Control Statements

if

documentation

The if statement has the following syntax:

if <condition> <then command> [else <command>]

There are three types of conditions:

  • <string 1>==<string 2>: most useful,
  • ERRORLEVEL <number>: test the return value of the last command, and
  • exists <file>: test if the file exists

Any of the conditions can be negated by using if not instead of if.

goto

documentation

The goto statement has the following syntax:

goto <label>

The <label> is then prese in the script as :<label>. Example:

goto label

...

:label
echo Hello, World!

Output redirection

unofficial documentation

The | operator is used to redirect the standard output of a command to the input of another command as usual.

The > is used to redirect the standard output of a command to a file. By prependind 2 (2>), we redirect the standard error output instead. When changing > to >>, we write to the file in the append mode.

Using < we can redirect a file to the standard input of a command.

Finally, 2>&1 redirects the standard error output to the standard output. This can be combined with the output to file: <command> > <file> 2>&1 writes both the standard output and the standard error output to the file.

As any <file> we can use NUL which basically discards the output.

Command Separation

We can execute multiple commands in the same line. We can use

  • <command 1> & <command 2>: commands are executed independently, or
  • <command 1> && <command 2>: <command 2> is executed only if <command 1> exits with code 0.

Batch Script Arguments

When we call commands from a batch script, we can refer to the called arguments called replacement parameters. These are %0 (the name of the script), %1--%9 (the arguments of the script). To refer all script arguments (%1--%9), we can use %*.

Contrary to a typical programming language behavior, the arguments are not split just by spaces but also commas (,), semi-colons (;), equal signs (=), and by a horizontal tab. Therefore, one has to be careful if some of these characters are used in the arguments.

Parameter extension for files

If the argument is a file name, we can get several additional information about the file by applying parameters extensions. These extensions are in format ~<extension> where <extension> is a single letter long and are placed between the % and the parameter number. Example:

echo %~f1

The above command prints the full path of the file passed as the first argument.

Most commonly used extensions are:

  • ~: remove any surrounding quotes
  • ~f: full path
  • ~d: drive letter
  • ~p: path, without the drive letter and the file name

We can also combine some extensions:

echo %~dp1

The above command prints the directory of the file passed as the first argument.

Usefull Commands

Get the path of the executable/command

To get the path of the executable/command, we can use the where command. Example:

where java

If there are more than one executable/command with the same name, all are listed (one per line), from the one to be executed, to the one with the lowest priority.

findstr: Select lines containing a pattern

The findstr command is used to select lines containing a pattern. It is similar to the grep or Select-String in PowerShell. However, it can only ever select the whole line, not parts of it.

The command can either:

  • read from a file: findstr <pattern> <file>, or
  • read from the standard input: <command> | findstr <pattern>.

By default, the pattern is interpreted as a regular expression. However, findstr has a plenty of options to configure the matching:

  • /i: ignore case
  • /c:<pattern>: use a literal pattern, instead of a regular expression

The mklink command is used to create a symlink. Example:

mklink <source> <target>

Limitations

The command shell is very limited compared to any other shell that is used in practice. Some of the limitations are:

  • no command for selecting line by a number like head/tail
  • no command for matching substrings, only whole line can be matched