Chapter 10
Functions and Arrays
Chapter 10: Shell Functions, Arrays, and String Processing
Functions are the foundation of script reuse, arrays enable bulk data processing, and advanced string operations let you say goodbye to awk/sed complexity. This chapter starts with function scope and return values, systematically covers indexed arrays, associative arrays, mapfile file reading, printf formatting, nameref, and closes with a complete production-grade function library.
10.1 Function Definition: Two Syntaxes
bash supports two function definition syntaxes that are functionally identical. The function keyword is a bash extension and not POSIX portable; the name() {} syntax works in all POSIX shells:
# 语法一：POSIX 兼容（推荐用于需要移植的脚本）
greet() {
echo "Hello, $1!"
}
# 语法二：bash 关键字语法（bash 专属）
function farewell {
echo "Goodbye, $1!"
}
# 两种语法可以混用，行为完全相同
greet "World" # → Hello, World!
farewell "World" # → Goodbye, World!
# 函数必须先定义再调用（bash 顺序执行）
# 例外：如果函数定义在 source 的文件中，调用时必须先 source
# 单行函数（用分号分隔语句）
log() { echo "[$(date +%T)] $*"; }
# 函数名可以包含连字符（bash 特性）
check-root() {
[[ $(id -u) -eq 0 ]] || { echo "需要 root 权限" >&2; return 1; }
}
# 查看已定义的函数
declare -F # 列出所有函数名
declare -f greet # 查看 greet 函数的完整定义
type greet # 显示 greet 是一个 function
# 删除函数
unset -f greet
10.2 Parameters and Return Values
Function parameters are accessed via positional variables $1~$9/${10}. return can only return an integer exit code (0-255). Two idiomatic patterns exist for returning string data:
#!/bin/bash
# === local 变量作用域 ===
# 函数内不声明 local 的变量是全局的！
x=10
modify_global() {
x=99 # 修改了全局变量 x
local y=42 # y 仅在函数内有效
}
modify_global
echo $x # → 99（全局被修改）
echo $y # → （空，y 在函数外不存在）
# 良好实践：函数内所有变量都用 local
calculate() {
local a="$1"
local b="$2"
local result=$(( a + b ))
echo $result # 通过 stdout 返回结果
}
# 捕获函数返回的字符串
sum=$(calculate 15 27)
echo "15 + 27 = $sum" # → 15 + 27 = 42
# === return 退出码 ===
is_even() {
local n="$1"
(( n % 2 == 0 )) # 算术表达式：0 为真(0)，非0 为假(1)
return $?
}
is_even 4 && echo "4 是偶数" || echo "4 是奇数"
is_even 7 && echo "7 是偶数" || echo "7 是奇数"
# === 通过全局变量返回复杂数据 ===
# 惯例：用 __ 前缀的全局变量作为"输出参数"
__split_result=()
split_string() {
local str="$1"
local sep="${2:- }"
IFS="$sep" read -r -a __split_result
## 10.3 Recursive Functions
```bash
#!/bin/bash
# 阶乘（递归）
factorial() {
local n="$1"
if (( n
## 10.4 Advanced Function Features
```bash
# 函数作为命令（覆盖外部命令）
# 注意：这会屏蔽同名的外部命令，使用时要谨慎
ls() {
command ls --color=auto -F "$@" # command 调用真正的 ls
}
# source 导入函数库
source /path/to/lib.sh # 在当前 Shell 中执行，共享作用域
# 简写
. /path/to/lib.sh
# 检查函数是否已定义
if declare -f my_func &>/dev/null; then
echo "my_func 已定义"
my_func
fi
# 列出所有函数
declare -F | awk '{print $3}' # 只打印函数名
# 删除函数
unset -f ls # 恢复 ls 为外部命令
# trap — 捕获信号，常用于函数中的清理操作
cleanup() {
echo "清理临时文件..."
rm -f /tmp/myapp.$$.*
}
trap cleanup EXIT # 脚本退出时自动调用 cleanup
trap cleanup INT # Ctrl+C 时也调用
# FUNCNAME 数组 — 函数调用栈
outer() {
inner
}
inner() {
echo "调用栈: ${FUNCNAME[*]}" # → inner outer main
echo "当前函数: ${FUNCNAME[0]}" # → inner
echo "调用者: ${FUNCNAME[1]}" # → outer
}
outer
10.5 Indexed Arrays (declare -a)
#!/bin/bash
# === 创建数组 ===
fruits=("apple" "banana" "cherry") # 直接赋值
declare -a colors=("red" "green" "blue") # 显式声明
nums=(1 2 3 4 5)
mixed=("hello" 42 "world" 3.14) # 类型混合（全为字符串）
# 按索引赋值（可以不连续）
arr[0]="first"
arr[1]="second"
arr[5]="sixth" # 中间空洞（3,4 不存在）
# === 访问元素 ===
echo ${fruits[0]} # → apple（第一个）
echo ${fruits[1]} # → banana
echo ${fruits[-1]} # → cherry（最后一个，bash 4.3+）
echo ${fruits[-2]} # → banana（倒数第二个）
# === 整体操作 ===
echo ${fruits[@]} # → apple banana cherry（所有元素）
echo ${fruits[*]} # → apple banana cherry（同上，在双引号内行为不同）
echo ${#fruits[@]} # → 3（元素个数）
echo ${!fruits[@]} # → 0 1 2（所有索引）
# $@ 与 $* 在引号内的区别（数组同理）
for item in "${fruits[@]}"; do echo "$item"; done # 每个元素独立
for item in "${fruits[*]}"; do echo "$item"; done # 合并为一个字符串
# === 增删改 ===
fruits+=("date" "elderberry") # 追加元素
fruits[1]="BANANA" # 修改元素
unset fruits[2] # 删除元素（留下空洞，索引不重排）
echo ${fruits[@]} # → apple BANANA date elderberry
# 删除整个数组
# unset fruits
# === 数组切片 ===
echo ${fruits[@]:1:2} # 从索引1开始取2个元素
# === 遍历（带索引）===
for i in "${!fruits[@]}"; do
echo "fruits[$i] = ${fruits[$i]}"
done
# === 数组排序 ===
unsorted=(banana apple cherry date elderberry)
IFS=$'\n' sorted=($(sort
## 10.6 Associative Arrays (declare -A) — bash 4+
```bash
#!/bin/bash
# 关联数组（字典），bash 4.0+ 才支持
# 检查版本：[[ ${BASH_VERSINFO[0]} -ge 4 ]] || { echo "需要 bash 4+"; exit 1; }
# === 创建关联数组 ===
declare -A user_info
user_info[name]="Alice"
user_info[age]="30"
user_info[role]="admin"
# 一次性初始化（bash 4+）
declare -A config=(
[host]="localhost"
[port]="5432"
[dbname]="myapp"
[user]="dbuser"
)
# === 访问 ===
echo ${user_info[name]} # → Alice
echo ${config[port]} # → 5432
# === 整体操作 ===
echo ${config[@]} # 所有值（顺序不保证）
echo ${!config[@]} # 所有键
echo ${#config[@]} # 键值对数量
# === 遍历键值对 ===
for key in "${!config[@]}"; do
echo "$key = ${config[$key]}"
done
# === 检查键是否存在 ===
key="host"
if [[ -v config[$key] ]]; then # -v 检查变量是否已设置（bash 4.2+）
echo "键 '$key' 存在: ${config[$key]}"
fi
# 兼容性写法（bash 4.0）
if [[ ${config[$key]+_} ]]; then
echo "键 '$key' 存在"
fi
# === 删除键 ===
unset config[user]
echo ${#config[@]} # → 3
# === 实战：统计单词频率 ===
declare -A word_count
text="the quick brown fox jumps over the lazy dog the fox"
for word in $text; do
(( word_count[$word]++ ))
done
# 按频率降序输出
for word in "${!word_count[@]}"; do
echo "${word_count[$word]} $word"
done | sort -rn
# === 实战：解析配置文件 ===
declare -A cfg
while IFS='=' read -r key value; do
[[ $key =~ ^[[:space:]]*# ]] && continue # 跳过注释
[[ -z $key ]] && continue # 跳过空行
cfg[${key// /}]="${value// /}" # 去除空格
done
## 10.7 mapfile / readarray: Reading Files into Arrays
```bash
#!/bin/bash
# mapfile（bash 4+，readarray 是其别名）
# 将文件的每一行读入数组，-t 去掉末尾换行符（几乎总是需要）
mapfile -t lines **Version Note:** Both `mapfile` and associative arrays (`declare -A`) require bash 4.0+. macOS ships with bash 3.2 (GPL licensing). Install a newer version with `brew install bash`, or check the version at the top of your script: `[[ ${BASH_VERSINFO[0]} -ge 4 ]] || exit 1`.
## 10.8 Formatted Output with printf
```bash
#!/bin/bash
# printf 比 echo 更可控，行为与 C 的 printf 一致
# 基本格式符
printf "%s\n" "hello" # 字符串
printf "%d\n" 42 # 整数
printf "%f\n" 3.14 # 浮点（默认6位小数）
printf "%.2f\n" 3.14159 # 保留2位小数
printf "%x\n" 255 # 十六进制（ff）
printf "%X\n" 255 # 十六进制大写（FF）
printf "%o\n" 8 # 八进制（10）
printf "%e\n" 12345.678 # 科学计数法
printf "%b\n" "hello\nworld" # 解释转义序列
# 宽度与对齐
printf "%-20s %5d\n" "Alice" 30 # 左对齐字符串，右对齐数字
printf "%-20s %5d\n" "Bob" 25
printf "%-20s %5d\n" "Charlie" 28
# 输出：
# Alice 30
# Bob 25
# Charlie 28
# 补零
printf "%05d\n" 42 # → 00042
printf "%08.2f\n" 3.14 # → 00003.14
# 多次重复格式（参数多于占位符时，格式会重复）
printf "%s\n" apple banana cherry
# → apple
# banana
# cherry
# 颜色输出（ANSI 转义码）
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
BLUE='\033[0;34m'
BOLD='\033[1m'
RESET='\033[0m'
printf "${RED}错误：${RESET}文件不存在\n"
printf "${GREEN}成功：${RESET}部署完成\n"
printf "${YELLOW}警告：${RESET}磁盘空间不足\n"
printf "${BOLD}${BLUE}信息：${RESET}服务已启动\n"
# 写入文件（不用 echo 和重定向）
printf "Name: %s\nAge: %d\n" "Alice" 30 > /tmp/user.txt
# 生成固定宽度的表格
printf "%-15s %-10s %-10s\n" "Name" "Status" "PID"
printf "%-15s %-10s %-10s\n" "nginx" "running" "1234"
printf "%-15s %-10s %-10s\n" "mysql" "stopped" "—"
# 进度条
progress_bar() {
local current="$1"
local total="$2"
local width=40
local pct=$(( current * 100 / total ))
local filled=$(( current * width / total ))
local empty=$(( width - filled ))
printf "\r[%-*s] %3d%%" $width "$(printf '#%.0s' $(seq 1 $filled))" $pct
[[ $current -eq $total ]] && echo
}
for i in $(seq 1 20); do
progress_bar $i 20
sleep 0.05
done
10.9 Advanced String Processing
#!/bin/bash
# === 正则匹配与 BASH_REMATCH ===
email="[email protected]"
if [[ $email =~ ^([a-zA-Z0-9._%+-]+)@([a-zA-Z0-9.-]+)\.([a-zA-Z]{2,})$ ]]; then
echo "合法邮箱"
echo "用户名: ${BASH_REMATCH[1]}" # → user
echo "域名: ${BASH_REMATCH[2]}" # → example
echo "顶级域: ${BASH_REMATCH[3]}" # → com
fi
# 提取 IP 地址
text="服务器地址：192.168.1.100，备用：10.0.0.1"
if [[ $text =~ ([0-9]+\.[0-9]+\.[0-9]+\.[0-9]+) ]]; then
echo "第一个 IP: ${BASH_REMATCH[1]}" # → 192.168.1.100
fi
# === 大小写转换（bash 4+）===
str="Hello World"
echo ${str,,} # → hello world（全小写）
echo ${str^^} # → HELLO WORLD（全大写）
echo ${str,} # → hELLO wORLD（仅首字母小写，若原为大写）
echo ${str^} # → Hello World（仅首字母大写）
# 逐字转换（指定字符）
echo ${str^^[aeiou]} # → HEllO WOrld（元音转大写）
# === 去除首尾空白（bash 无内置，用参数展开模拟）===
trim() {
local str="$1"
# 去除前导空白
str="${str#"${str%%[![:space:]]*}"}"
# 去除尾部空白
str="${str%"${str##*[![:space:]]}"}"
echo "$str"
}
result=$(trim " hello world ")
echo "[$result]" # → [hello world]
# 更简洁的写法（使用正则）
trim2() {
local str="$*"
[[ $str =~ ^[[:space:]]*(.*[^[:space:]])[[:space:]]*$ ]] && echo "${BASH_REMATCH[1]}" || echo ""
}
# === 字符串重复 ===
repeat() {
local str="$1"
local n="$2"
printf "%${n}s" | tr ' ' "$str"
}
repeat '-' 40 # → ----------------------------------------
# === 字符串包含检测 ===
haystack="Hello, World!"
needle="World"
if [[ $haystack == *"$needle"* ]]; then
echo "包含 '$needle'"
fi
# === 字符串分割 ===
csv="a,b,c,d,e"
IFS=',' read -r -a parts
## 10.10 Name References (declare -n)
`declare -n` (nameref, bash 4.3+) creates a variable whose value is the name of another variable. Through nameref, functions can return arrays and other complex data structures:
```bash
#!/bin/bash
# 基本 nameref
real_var="Hello"
declare -n alias=real_var # alias 是 real_var 的别名
echo "$alias" # → Hello（访问 real_var 的值）
alias="World" # 通过 alias 修改 real_var
echo "$real_var" # → World
# === 函数通过 nameref 返回数组 ===
# 这是 bash 返回数组的最佳实践（bash 4.3+）
get_users() {
declare -n _result="$1" # $1 是调用者传入的数组名
_result=()
while IFS=: read -r name _ uid _; do
(( uid >= 1000 )) && _result+=("$name")
done
## 10.11 Function Library Design: Complete Example
Below is a complete production-grade shell function library, including logging, colored output, error handling, and duplicate-load prevention:
```bash
#!/bin/bash
# lib/common.sh — 通用函数库
# 使用方法：source lib/common.sh
# ============================================================
# 防重复加载（幂等 source）
# ============================================================
[[ -n "${__COMMON_LIB_LOADED:-}" ]] && return 0
readonly __COMMON_LIB_LOADED=1
# ============================================================
# 颜色常量
# ============================================================
readonly _CLR_RESET='\033[0m'
readonly _CLR_RED='\033[0;31m'
readonly _CLR_GREEN='\033[0;32m'
readonly _CLR_YELLOW='\033[1;33m'
readonly _CLR_BLUE='\033[0;34m'
readonly _CLR_CYAN='\033[0;36m'
readonly _CLR_BOLD='\033[1m'
# ============================================================
# 日志函数
# ============================================================
# 内部：带时间戳输出
_log() {
local level="$1"
local color="$2"
shift 2
local msg="$*"
local timestamp
timestamp=$(date '+%Y-%m-%d %H:%M:%S')
printf "${color}[${timestamp}] [${level}]${_CLR_RESET} %s\n" "$msg" >&2
}
# 公共日志函数
log_info() { _log "INFO " "${_CLR_GREEN}" "$@"; }
log_warn() { _log "WARN " "${_CLR_YELLOW}" "$@"; }
log_error() { _log "ERROR" "${_CLR_RED}" "$@"; }
log_debug() {
[[ "${LOG_LEVEL:-info}" == "debug" ]] || return 0
_log "DEBUG" "${_CLR_CYAN}" "$@"
}
# 致命错误：打印后退出
log_fatal() {
_log "FATAL" "${_CLR_RED}${_CLR_BOLD}" "$@"
exit 1
}
# ============================================================
# 错误处理
# ============================================================
# 设置错误处理模式（在主脚本调用）
enable_strict_mode() {
set -euo pipefail
trap '_handle_error ${LINENO} "$BASH_COMMAND"' ERR
}
_handle_error() {
local line="$1"
local cmd="$2"
log_error "命令失败（第 ${line} 行）: ${cmd}"
log_error "退出码: $?"
}
# ============================================================
# 工具函数
# ============================================================
# 检查命令是否存在
require_cmd() {
local cmd
for cmd in "$@"; do
command -v "$cmd" &>/dev/null || log_fatal "必须安装命令: ${cmd}"
done
}
# 检查是否以 root 运行
require_root() {
[[ $(id -u) -eq 0 ]] || log_fatal "此脚本必须以 root 运行"
}
# 安全创建目录
ensure_dir() {
local dir="$1"
local mode="${2:-755}"
if [[ ! -d "$dir" ]]; then
mkdir -p "$dir" && chmod "$mode" "$dir" || log_fatal "无法创建目录: ${dir}"
log_info "创建目录: ${dir}"
fi
}
# 确认提示（返回 0 确认，1 取消）
confirm() {
local msg="${1:-确认继续？}"
local answer
read -r -p "${msg} [y/N] " answer
[[ ${answer,,} == "y" ]]
}
# 重试函数（最多 N 次）
retry() {
local max_attempts="${1:-3}"
local delay="${2:-2}"
shift 2
local attempt=1
while (( attempt backup.sql
# log_info "备份完成：$(format_size $(stat -c%s backup.sql))"
# fi
#
# log_info "部署完成"
Function Library Best Practices: A good shell function library should have: (1) duplicate-load guard (
__LIB_LOADED); (2)localfor all internal variables; (3) private functions prefixed with_; (4) consistent timestamped log format; (5) all errors written to stderr (>&2); (6) anenable_strict_modehelper for callers to opt in to strict mode.
Chapter Summary: This chapter fully covered the bash function system (definition/scope/recursion/nameref), all operations on indexed and associative arrays, efficient file reading with mapfile, printf formatting and color output, advanced string processing, and a production-ready function library template. The next chapter dives into pipes, process substitution, and the underlying mechanics of file descriptors.
Previous
← Ch9: Variables
Next
Ch11: Pipes →