How to Write Shell Scripts

Posted on 2023-04-23 Edited on 2026-09-09 In Computer Systems Views:

The basic grammer of shell/Bash script.

ref: shell tutorial

How to Write Shell Script

首先做个澄清：Shell 是一类程序的统称，其中最流行的是 Bash Shell，所有 Linux 发行版和 MacOS 都安装了它；除此之外，zsh 也很流行。考虑到 Bash 的流行程度以及其他更复杂的 Shell 的功能基本是 Bash 的超集，因此我们这里只介绍 Bash Shell 脚本。

前置知识：Shell 是如何执行命令的

Shell 在执行命令时，通常先 fork 出一个子进程，再由子进程 exec 目标程序。

这里有两个常见例外：

Shell 内建（built-in）命令，例如 cd、export，通常直接在当前 Shell 进程中执行。（但不包括 ls，ls不是内建命令）
使用 source 执行脚本，例如 source any-script.sh，脚本内容会直接在当前 Shell 中执行。

Shell 中的变量可以分为普通 Shell 变量和环境变量。普通变量不会被子进程继承（或者说复制）；环境变量会被复制给子进程。这也意味着子进程修改自己的环境变量不会影响父进程。

定义方式如下：

# 普通变量
FOO=hello

# 环境变量
export FOO=hello

前置知识：执行shell脚本的具体方式

执行脚本文件：

直接指定sh来执行该脚本，不需要shebang，也不需要脚本有执行权限（因为该脚本直接作为参数传给了sh).
sh script_file_path
需要写shebang来指定解释器, 并且要指定脚本路径.

必须加上./使得该command name被识别为一个路径名. 否则shell会继续在alias, builtin和PATH中搜索该command name. ¹
chmod +x script_file ##(chown, chgrp optionally)
./script_file
使用.或source在当前shell session中执行该脚本
source script_file
or
. script_file
因此, 该方法可以用于刷新当前shell环境:
source ~/.bashrc
Shell script是能在命令行直接输入的, 但仅会作用一次

注意: 方法1,2都是新开一个子shell session，在其中执行脚本，而方法3, 4是在当前shell session中执行脚本

Environment virable

$PATH详见Shell Script Searching Path

常见的环境变量:

环境变量	说明
$HOME	当前用户的登陆目录
$PATH	以冒号分隔的, 由多个路径所组成的, 用来搜索命令的列表
$PS1	命令行提示符，通常是”$”字符 (很多主题都会改掉$PS1)
$PS2	辅助提示符，用来提示后续输入，通常是”>”字符
$IFS	输入区分隔符。当shell读取输入数据时会把一组字符看成是单词之间的分隔符，通常是空格、制表符、换行符等

#我们在当前Shell进程中指定了var1变量
[roc@roclinux ~]$ var1="hello"
[roc@roclinux ~]$ echo $var1
hello

#我们在当前Shell进程中又指定了var2变量
[roc@roclinux ~]$ var2="world"
[roc@roclinux ~]$ echo $var2
world

#我们通过export发布var1
[roc@roclinux ~]$ export var1
 
#我们进入到一个bash子进程中
[roc@roclinux ~]$ bash
 
#var1变量仍然有效, 而var2变量已经无效了
[roc@roclinux ~]$ echo $var1 # 输出变量的值
hello
[roc@roclinux ~]$ echo $var2  # 什么也没有输出

例子：用 source 执行会更改环境变量的脚本

假设有一个脚本 reset_env.sh：

export FOO=new_value

如果直接执行：

export FOO=old_value
./reset_env.sh

echo "$FOO"

输出仍然是：

old_value

因为 reset_env.sh 在子进程中执行，修改的是子进程自己的环境。

如果改成：

source reset_env.sh

echo "$FOO"

输出则是：

new_value

因为 source 会在当前 Shell 中执行脚本，所以能够直接修改当前 Shell 的环境变量。

Parameter variable

$0 - Name of the script
$1 to $9 - Arguments to the script. $1 is the first argument and so on.
- 当n>=10时，需要使用${n}来获取参数
$@ - 全部参数组成的列表
$# - Number of arguments
$? - Return code of the previous command
$*：全部参数连接成的字符串，按$IFS的第一个字符分割
$$ - Process identification number (PID) for the current script
!! - Entire last command, including arguments. A common pattern is to execute a command only for it to fail due to missing permissions; you can quickly re-execute the command with sudo by doing sudo !!
$_ - Last argument from the last command. If you are in an interactive shell, you can also quickly get this value by typing Esc followed by . or Alt+.

Variable

assign variables in bash: foo=bar
- Note that foo = bar will not work since it is interpreted as calling the foo program with arguments = and bar. In general, in shell scripts the space character will perform argument splitting.
  - 因此不要加空格
access the value of the variable: $foo
- 等价于 ${foo}，花括号可以精确地界定变量名称的范围。

可以用read命令从标准输入接受数据并赋值:read val

example:

#! /usr/bin/env bash
echo -n "Enter your name:"
read name
echo "hello $name"
exit 0

String

Strings in bash can be defined with ' and " delimiters, but they are not equivalent.
Strings delimited with ' are literal strings and will not substitute variable values whereas " delimited strings will.
foo=bar
echo "$foo"
# prints bar
echo '$foo'
# prints $foo

Array

array initialization −
array_name=(value1 ... valuen)
assign
array_name[index]=value
access Array Values
${array_name[index]}

Quoting Mechanism

metacharacters

Unix Shell provides various metacharacters which have special meaning while using them in any Shell Script and causes termination of a word unless quoted.

* ? [ ] ' " \ $ ; & ( ) | ^ < > new-line space tab

quoting

The following table lists the four forms of quoting −

Sr.No.	Quoting & Description
1	Single quote All special characters between these quotes lose their special meaning.
2	Double quote Most special characters between these quotes lose their special meaning with these exceptions −$`$'"\
3	Backslash Any character immediately following the backslash loses its special meaning.
4	Back quote (aka backtick) Everything you type between backticks is evaluated (executed) by the shell before the main command, and the output of that execution is used by that command,

The Single Quotes: 其内容不转义, 相当于在每个字符前加 backslash
The Double Quotes：其内容转义
The Backslash: 取消其后面的一个字符的转义
The Backquotes：将其内容视作 command 并执行，与后文的CMD substitution类似
var=`command`
DATE=`date`

echo $DATE

#等价于后文的CMD substitution
echo $(DATE)

常见语法

条件语句

As with most programming languages, bash supports control flow techniques including if, case, while and for.

if

syntax:

if [ expression ] 
then 
   Statement(s) to be executed if expression is true 
fi

if [ expression 1 ]
then
   Statement(s) to be executed if expression 1 is true
elif [ expression 2 ]
then
   Statement(s) to be executed if expression 2 is true
elif [ expression 3 ]
then
   Statement(s) to be executed if expression 3 is true
else
   Statement(s) to be executed if no expression is true
fi

紧凑形式： ; (同一行上多个命令的分隔符)

example:

if [ -f ~/.bashrc ]; then 
. ~/.bashrc fi

#!/bin/sh
read -p "Is this morning? Please answer yes or no: " answer
if [ "$answer" = "yes" ]; then
 echo “Good morning”
elif [ "$answer" = "no" ]; then
 echo “Good afternoon” 
else
 echo “Sorry, $answer not recognized. Enter yes or no”
 exit 1 
fi
exit 0

case

syntax:

case word in
   pattern1 | pattern2)
      Statement(s) to be executed if pattern1 matches
      ;;
   pattern3)
      Statement(s) to be executed if pattern2 matches
      ;;
   pattern4)
      Statement(s) to be executed if pattern3 matches
      ;;
   *)
     Default condition to be executed
     ;;
esac

example:

#!/bin/sh

option="${1}" 
case ${option} in 
   -f) FILE="${2}" 
      echo "File name is $FILE"
      ;; 
   -d) DIR="${2}" 
      echo "Dir name is $DIR"
      ;; 
   *)  
      echo "`basename ${0}`:usage: [-f file] | [-d directory]" 
      exit 1 # Command to come out of the program with status 1
      ;; 
esac

#!/bin/sh

read -p "Is this morning? Please answer yes or no." answer
case "$answer" in
    yes | y | Yes | YES) 
        echo “Good morning!” 
        ;; 
    no | n | No | NO) 
        echo “Good afternoon!” 
        ;; 
    *)
        echo “Sorry, $answer not recognized.” 
        ;;
esac 
exit 0

select in && case in

select in语句自带循环

select variable in value_list
do
  statements
done

variable: 表示变量
value_list: 取值列表
in: Shell关键字

select in 通常和 case in 一起使用，在用户输入不同的编号时可以做出不同的反应

example:

#!/bin/sh
clear
select item in Continue Finish
do
    case "$item" in
    Continue) 
        ;;
    Finish) 
        break 
        ;;
    *) 
        echo "Wrong choice! Please select again!" 
        ;;
    esac 
done

该命令的while版本：

while [ "$item" != "Finish" ]; 
do 
    read item
    case "$item" in
        "Continue") 
            ;;
        "Finish")
            ;;
        *) echo "Wrong choice! Please select again!"
            ;;
    esac 
done

循环语句

break: 从for/while/until循环退出

for

syntax：

for var in word1 word2 ... wordN
do
   Statement(s) to be executed for every word.
done

example:

for FILE in $HOME/.bash*
do
   echo $FILE
done

for f in *.png             
do
	mv -n $f $f.kk
done

while

syntax:

while condition 
do
	statements 
done

example:

quit=n
while [ "$quit" != "y" ]; do read menu_choice
case "$menu_choice" in
a) echo "a, continue...";;
b) echo "b, continue...";;
q|Q) quit=y;;
*) echo "Sorry, choice not recognized.";;
esac done

a=0
LIMIT=10
while [ $a -le $LIMIT ] 
do
    a=$(($a+1))
        if [ $a -gt 2 ] 
        then
            break # Skip entire rest of loop. 
        fi
    echo -n "$a"
done

命令组合

分号串联：

command1 ; command2 ; ...

条件组合:

statement1 && statement2 && statement3 && ...

statement1 || statement2 || statement3 ||

语句块

{
	statement1 
	statement2 
	...
}

或

{ 
	statement1; statement2 ; ... ; 
}

test expression 或 `[ expression ]`

[ (aka test) command the and [[ ... ]] test construct are used to evaluate expressions
[是一条命令，与test等价，大多数shell都支持。在现代的大多数sh实现中，[与test是builtin命令
- []将其operand直接当作argument
[[，是关键字，许多shell(如ash bsh)并不支持这种方式
- [[]]将其operand进行参数引用，算术扩展和CMD substitution，不需要手动转义等

Function

syntax

function_name(){
	list of commands
	[ return value ]
}

example:

mcd () {
    mkdir -p "$1"
    cd "$1"
}

return

exit: 不仅会退出函数, 还会退出执行该函数的shell
return code:仅仅退出函数. 和command的return code同

return code

Commands have Return Code to report errors in a more script-friendly manner.
- 0 usually means everything went OK; anything different from 0 means an error occurred.
- return code，你可以指定返回任何值

return code as bool value：

Return codes can be used to conditionally execute commands using && (and operator) and || (or operator) .
Commands can also be separated within the same line using a semicolon ;.

The true program will always have a 0 return code and the false command will always have a 1 return code.

false || echo "Oops, fail"
# Oops, fail

true || echo "Will not be printed"
#

true && echo "Things went well"
# Things went well

false && echo "Will not be printed"
#

true ; echo "This will always run"
# This will always run

false ; echo "This will always run"
# This will always run

Shebang

Note that scripts need not necessarily be written in bash to be called from the terminal. For instance, here’s a simple Python script that outputs its arguments in reversed order:

#!/usr/local/bin/python
import sys
for arg in reversed(sys.argv[1:]):
    print(arg)

shebang: the character sequence consisting of #! at the beginning of a script in a Unix-like operating system
- shell会将shebang中#!之后的内容作为一个程序的路径，打开该程序，将本script的路径当作参数传入( 即: 将整个script当作input传入shebang所指定的程序）
- For example, if a script is named with the path path/to/script, and it starts with the following line, #!/bin/sh, then the program loader is instructed to run the program /bin/sh, passing path/to/script as the first argument. In Linux, this behavior is the result of both kernel and user-space code.[9]
- The shebang line is usually ignored by the interpreter, because the "#" character is a comment marker in many scripting languages; some language interpreters that do not use the hash mark to begin comments still may ignore the shebang line in recognition of its purpose.

she-bang with env

The she-bang expects a full path to the interpreter to use so the following syntax would be incorrect:

#!python

Setting a full path like this might work:

#!/usr/local/bin/python

but would be non portable as python might be installed in /bin, /opt/python/bin, or wherever other location.

Using env

#!/usr/bin/env python

is a method allowing a portable way to specify to the OS a full path equivalent to the one where python is first located in the PATH.

I/O redirection

一般情况下，每个 Linux 命令运行时都会打开三个文件：

标准输入文件(stdin)：stdin的文件描述符为0，Unix程序默认从stdin读取数据。
标准输出文件(stdout)：stdout 的文件描述符为1，Unix程序默认向stdout输出数据。
标准错误文件(stderr)：stderr的文件描述符为2，Unix程序会向stderr流中写入错误信息
Discard the output
$ command > /dev/null
The file /dev/null is a special file that automatically discards all its input.
Discard both output of a command and its error output,
$ command > /dev/null 2>&1
- a command normally writes its output to STDOUT
- use standard redirection to redirect STDERR to STDOUT
- 这里2>&1将标准错误（2）合并到标准输出（1），而标准输出已经被重定向到了/dev/null, 因此总体效果是，标准错误和输出都被重定向到了/dev/null
file descriptor:
- 0 : STDIN
- 1: STDOUT
- 2: STDERR

Sr.No.	Command & Description
1	pgm > file Output of pgm is redirected to file 会覆盖目标文件中原有的数据
2	pgm < file Program pgm reads its input from file
3	pgm >> file Output of pgm is appended to file
4	n > file Output from stream with descriptor n redirected to file
5	n >> file Output from stream with descriptor n appended to file
6	n >& m Merges output from stream n with stream m
7	n <& m Merges input from stream n with stream m
8	<< tag Standard input comes from here through next tag at the start of line
9	\| Takes output from one program, or process, and sends it to another

Here Document

Here Document 目前没有统一的翻译，这里暂译为”嵌入文档“。Here Document 是 Shell 中的一种特殊的重定向方式，它的基本的形式如下：

command << delimiter
document
delimiter

它的作用是将两个 delimiter 之间的内容(document) 作为输入传递给 command

注意：

结尾的delimiter 一定要顶格写，前面不能有任何字符，后面也不能有任何字符，包括空格和 tab 缩进。
开始的delimiter前后的空格会被忽略掉。

下面的例子，通过 wc -l 命令计算 document 的行数：

wc -l << EOF
    This is a simple lookup program
    for good (and bad) restaurants
    in Cape Town.
EOF

也可将 Here Document 用在脚本中，例如：

#!/bin/bash

cat << EOF
    This is a simple lookup program
    for good (and bad) restaurants
    in Cape Town.
EOF

https://lyk-love.cn/2023/04/24/Shell-Script-Searching-Path/↩︎

How to Write Shell Script

前置知识：Shell 是如何执行命令的

前置知识：执行shell脚本的具体方式

Environment virable

例子：用 source 执行会更改环境变量的脚本

Parameter variable

Variable

String

Array

Quoting Mechanism

metacharacters

quoting

常见语法

条件语句

if

case

select in && case in

循环语句

for

while

命令组合

语句块

test expression 或 [ expression ]

Function

return

return code

Shebang

she-bang with env

I/O redirection

Here Document

test expression 或 `[ expression ]`