主页 / coding / 问题

问题[bash](coding)

Martin Hope
Andres Mora
Asked: 2025-01-01 03:08:55 +0800 CST
  • 6

我需要根据以下情况删除 JPG 和 jpg 文件

文件夹有多个 JPG 和 jpg 文件。每个文件的名称例如：172.30.165.212_20241231_132125.JPG。其中172.30.165.212是 IP 地址，20241231是 YYYYMMDD 格式的日期，132125是 HHMMSS 格式的时间。

删除条件如下：
1- 脚本应始终根据文件名的日期/时间保留每个 IP 地址的最新文件。无论日期/时间有多旧。2-
但是，由于每个 IP 地址可以有多个文件，因此脚本应删除文件名中的日期/时间比当前时间早 2 小时以上的所有文件。3-
永远不要查看文件的修改日期/时间，只查看名称中的修改日期/时间。

我尝试过这个，但没有成功，因为文件没有被删除。

#!/bin/bash
# Define target directory
TARGET_DIR="/mnt/moe/results"
# Log start time
echo "[$(date)] Starting cleanup process in $TARGET_DIR"

# Function to process files for a single IP
process_ip_files() {
    local ip_prefix=$1
    local ip_files
    # Find files matching the IP
    ip_files=$(find "$TARGET_DIR" -type f -iname "${ip_prefix}_*" | sort)

    # Skip if no files
    if [[ -z "$ip_files" ]]; then
        echo "No files found for IP: $ip_prefix"
        return
    fi

    echo "Processing IP: $ip_prefix"

    # Variables to track files and the most recent file
    local most_recent_file=""
    local most_recent_time=0
    local files_to_delete=()

    # Get current time in seconds since epoch
    current_time=$(date +%s)

    # Iterate over files to determine the most recent and deletion criteria
    while IFS= read -r file; do
        echo "Processing file: $file"

        # Remove the path and get just the file name
        base_file=$(basename "$file")
        echo "Base file name: $base_file"

        # Split file name into components
        IFS='_' read -r ip date time ext <<< "$base_file"
        
        # Validate the expected number of fields and format
        if [[ -z "$ip" || -z "$date" || -z "$time" || "$ext" != "JPG" && "$ext" != "jpg" ]]; then
            echo "  Skipping file (does not match expected format): $file"
            continue
        fi

        # Check the timestamp format (YYYYMMDD HHMMSS)
        if ! [[ "$date" =~ ^[0-9]{8}$ ]] || ! [[ "$time" =~ ^[0-9]{6}$ ]]; then
            echo "  Skipping file (invalid timestamp format): $file"
            continue
        fi

        # Convert to seconds since epoch
        timestamp="$date $time"
        file_time=$(date -d "$timestamp" +%s)

        echo "  File: $file"
        echo "    Timestamp: $timestamp"
        echo "    File time (epoch): $file_time"
        echo "    Current time (epoch): $current_time"

        # Check if this file is the most recent one for the IP
        if (( file_time > most_recent_time )); then
            # If we already have a most recent file, we add it to the delete list
            if [[ -n "$most_recent_file" ]]; then
                files_to_delete+=("$most_recent_file")
            fi
            most_recent_file="$file"
            most_recent_time="$file_time"
        else
            # Check if the file is older than 2 hours (7200 seconds)
            if (( current_time - file_time > 7200 )); then
                echo "    Marking for deletion: $file"
                files_to_delete+=("$file")
            fi
        fi
    done <<< "$ip_files"

    # Display the most recent file for this IP
    echo "Most recent file for IP $ip_prefix: $most_recent_file"

    # Deleting files not the most recent one
    if [[ ${#files_to_delete[@]} -gt 0 ]]; then
        echo "Files marked for deletion for IP $ip_prefix:"
        for file in "${files_to_delete[@]}"; do
            echo "  - $file"
        done

        for file in "${files_to_delete[@]}"; do
            if [[ "$file" != "$most_recent_file" ]]; then
                echo "Deleting file: $file"
                rm -v "$file"
            fi
        done
    else
        echo "No files to delete for IP $ip_prefix."
    fi
}

# Process unique IP addresses
find "$TARGET_DIR" -type f \( -iname "*.jpg" -o -iname "*.JPG" \) -printf "%f\n" | \
    awk -F'_' '{print $1}' | sort -u | while read -r ip; do
    process_ip_files "$ip"
done

# Log completion
echo "[$(date)] Cleanup process finished."

例如，我有以下文件，当前日期/时间是 20241231 13:30

172.30.165.212_20241231_132125.JPG  
172.30.165.212_20241231_122125.JPG  
172.30.165.212_20241231_112125.JPG  
172.30.165.212_20241231_102125.JPG  
172.30.165.212_20241231_092125.JPG  
172.30.165.213_20241231_062125.JPG  
172.30.165.213_20241231_032125.JPG  
172.30.165.213_20241231_012125.JPG

脚本应该删除

172.30.165.212_20241231_112125.JPG (older than 2 hours)  
172.30.165.212_20241231_102125.JPG (older than 2 hours)  
172.30.165.212_20241231_092125.JPG (older than 2 hours)  
172.30.165.213_20241231_032125.JPG (older than 2 hours)  
172.30.165.213_20241231_012125.JPG (older than 2 hours)

脚本应该保留

172.30.165.212_20241231_132125.JPG  (younger than 2 hours)  
172.30.165.212_20241231_122125.JPG  (younger than 2 hours)  
172.30.165.213_20241231_062125.JPG  (older than 2 hours but most recent from this ip address)
bash
Martin Hope
NanoTech
Asked: 2024-12-30 18:05:41 +0800 CST
  • 6

我正在尝试获取某些日志的日期，以某种格式进行一些比较，这是我的命令：

fgrep "<expression>" <logFile> | sort | awk -F "[][]" 'messageDate=$(date -d "$2" "+%Y.%j.%H.%M.%S") { print messageDate }'

我打印的是日志文件的完整行。

如果我运行这个：

fgrep "<expression>" <logFiles> | sort | awk -F "[][]" 'messageDate=$(date -d "$2" "+%Y.%j.%H.%M.%S") { print $2 }'

我得到了日期但不是我想要的格式。

2014-09-04T08:22:16Z
2017-10-08T16:05:06Z
2022-11-30T14:50:16Z

日志文件包含如下消息：

[2022-11-30T14:50:16Z] <Info/Warning/Error>: <log message>

有谁明白为什么 awk 可以工作，这意味着它正确地分割文件，但是代码messageDate=$(date -d "$2" "+%Y.%j.%H.%M.%S")以某种方式再次获取完整的日志消息？

bash
Martin Hope
Roger Costello
Asked: 2024-12-29 20:32:54 +0800 CST
  • 8

每个文件的第一行包含字段名称。字段名称中可能存在重复。我只想打印唯一的字段名称。以下是我尝试的：

在 Bash 文件 files_and_folders.sh 中，我输入了以下内容：

#!/bin/bash
for file in **/*.TXT ; do
   awk 'NR == 1 { for (i=1; i<=NF; i++) if (!seen[$i]) seen[$i] = 1} END { for (idx in seen) printf ("%s\n",idx) }' "${file}"
done

Bash 文件成功运行，但输出包含重复项：

AB_CODE
ACFT_CODE
AC_TYPE
ADD_INFO
AKA
ALT
ALT
ALT
ALT
ALT
ALT
ALT
ALT1_DESC
ALT2_DESC
ALT3_DESC

如何修改 AWK 程序（在 Bash 脚本中）以消除重复？

bash
Martin Hope
pwp
Asked: 2024-12-29 17:52:49 +0800 CST
  • 5

我想在 bash 下使用 printf 来格式化德语动词，如下所示：

INFINITIV       PRATERITUM      PERFEKT         3.PERS. SING
heißen          hieß            hat geheißen    er heißt
kaufen          kaufte          hat gekauft     er kauft

我在我的 bash 脚本中使用了一个简单的 printf 命令：

printf "%-15s %-15s %-15s %-15s %-15s\n" "INFINITIV" "PRATERITUM" "PERFEKT" "3.PERS. SING"
printf "%-15s %-15s %-15s %-15s %-15s\n" "heißen" "hieß" "hat geheißen" "er heißt"
printf "%-15s %-15s %-15s %-15s %-15s\n" "kaufen" "kaufte" "hat gekauft" "er kauft"

但似乎 Eszett 被视为两个字符。因此，屏幕上每次出现 ß 后，文本都会向左移动一个位置。输出为：

pwp@linuxcomputers:~/Scripts/Test$ Deutsche-Verben-Scharf-S 
INFINITIV       PRATERITUM      PERFEKT         3.PERS. SING                   
heißen         hieß           hat geheißen   er heißt                      
kaufen          kaufte          hat gekauft     er kauft

我怎样才能防止这种转变并使 printf 格式按照预期进行？

bash
Martin Hope
gcb
Asked: 2024-12-26 07:16:13 +0800 CST
  • 6

看到一些让我怀疑自己理智的事情：

#!/bin/env bash
set -e
for d in '/tmp/somedir/*'; do               
    for f in "${d}/*pub"; do
        echo $f;
    done
done

它按预期返回

/tmp/somedir/AAA/fileaa.pub /tmp/somedir/BBB/filebb.pub

但是，如果我改为echo $f或echo "$f"我echo "${f}"得到：

/tmp/somedir/*/*pub

我很难理解为什么。首先，这是一个单独的项目，所以它不只是影响生产echo线的东西。

使用GNU bash, version 5.2.37(1)-release (x86_64-pc-linux-gnu)

这是该主机上的商店

assoc_expand_once       off
cdable_vars     off
cdspell         off
checkhash       off
checkjobs       off
checkwinsize    on
cmdhist         on
compat31        off
compat32        off
compat40        off
compat41        off
compat42        off
compat43        off
compat44        off
complete_fullquote      on
direxpand       off
dirspell        off
dotglob         off
execfail        off
expand_aliases  on
extdebug        off
extglob         on
extquote        on
failglob        off
force_fignore   on
globasciiranges on
globskipdots    on
globstar        off
gnu_errfmt      off
histappend      on
histreedit      off
histverify      off
hostcomplete    off
huponexit       off
inherit_errexit off
interactive_comments    on
lastpipe        off
lithist         off
localvar_inherit        off
localvar_unset  off
login_shell     off
mailwarn        off
no_empty_cmd_completion off
nocaseglob      off
nocasematch     off
noexpand_translation    off
nullglob        off
patsub_replacement      on
progcomp        on
progcomp_alias  off
promptvars      on
restricted_shell        off
shift_verbose   off
sourcepath      on
varredir_close  off
xpg_echo        off
bash
Martin Hope
konchy
Asked: 2024-12-24 15:56:57 +0800 CST
  • 6

我是 Bash 脚本编程的新手，我正在尝试理解以下代码：

tmp_file=/tmp/tmp_file$$
mkfifo $tmp_file
echo "msg_A" >$tmp_file # blocks, since pipe is unbuffered and no one is reading from it
read msg <$tmp_file
echo $msg

tmp_file=/tmp/tmp_file$$
mkfifo $tmp_file

exec 7<>$tmp_file # add this line

echo "msg_A" >$tmp_file # now, the write operation won't block, why?
read msg <$tmp_file
echo $msg  # msg_A is printed

我想知道exec 7<>$tmp_file上面的代码示例中做了什么，以及为什么添加此行会使写入操作非阻塞？

bash
Martin Hope
Hemant Kumar
Asked: 2024-12-23 22:13:13 +0800 CST
  • 5

我有一个使用 yq 读取的文件，并且必须使用 kustomization 模板中的 yaml 格式的配置。因此，我从文件读取配置的方式如下：

istio_ns_scoping_labels=$(yq -r '.cluster.config.istio_ns_scoping_labels' "$config_file" 2>/dev/null)
echo "$istio_ns_scoping_labels"

#istio_ns_scoping_labels
    [
  {
    "istio-discovery": "enabled"
  },
  {
    "kubernetes.io/metadata.name": "keycloak-idc-identity-broker"
  },
  {
    "kubernetes.io/metadata.name": "keycloak-idc-identity-broker-2"
  },
  {
    "kubernetes.io/metadata.name": "keycloak-idc-identity-broke-3"
  }
]

现在我想将其格式化为 yaml，以便可以将其替换为 yaml 模板，如下所示

      discoverySelectors:
        - matchLabels:
            istio-discovery: enabled
        - matchLabels:
            kubernetes.io/metadata.name: keycloak-idc-identity-broker
        - matchLabels:
            kubernetes.io/metadata.name: keycloak-idc-identity-broker-2
        - matchLabels:
            kubernetes.io/metadata.name: keycloak-idc-identity-broker-3

我必须确保 $labels 格式正确为 yaml 格式，因为我使用 envsubst 来替换模板内部的块，如下所示

    meshConfig:
      accessLogFile: /dev/stdout
      discoverySelectors:
        - matchLabels:
          ${MATCH_LABELS_BLOCK}

我正在使用 yq 从 config.yml 文件中读取标签配置，如下所示

config_file=cluster.yml
istio_ns_scoping_labels=$(yq -r '.cluster.config.istio_ns_scoping_labels' "$config_file" 2>/dev/null)

#cluster.yml
cluster:
  config:
    istio_ns_scoping_labels: 
      istio-discovery: enabled
      kubernetes.io/metadata.name: keycloak-idc-identity-broker
bash
Martin Hope
aenagy
Asked: 2024-12-23 08:21:05 +0800 CST
  • 6

这是 awk gsub 的延续，没有替换字段中所有句点实例。

我有一些意见：

$ cat ./ipv4settings.txt
ipv4.gateway:                           192.168.2.1
ipv4.routes:                            --
ipv4.route-metric:                      -1
ipv4.route-table:                       0 (unspec)

我能够生成所需的输出（某种程度上）：

$ awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print $1"="$2 }' ./ipv4settings.txt
ipv4_gateway=192.168.2.1
ipv4_routes=--
ipv4_route_metric=-1
ipv4_route_table=0 (unspec)

接下来我要为每一行输出声明变量。我使用了几种变体declare：

$awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print $1"="$2 }' ./ipv4settings.txt
ipv4_gateway=192.168.2.1
ipv4_routes=--
ipv4_route_metric=-1
ipv4_route_table=0 (unspec)
$ declare $( awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print $1"="$2 }' ./ipv4settings.txt )
-bash: declare: `(unspec)': not a valid identifier

awk我尝试引用（declare not a valid identifier bash ）的整行输出（https://www.baeldung.com/linux/awk-print-quote-characters）：

$ awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print "\042"$1"="$2"\042" }' ./ipv4settings.txt
"ipv4_gateway=192.168.2.1"
"ipv4_routes=--"
"ipv4_route_metric=-1"
"ipv4_route_table=0 (unspec)"
$ declare $( awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print "\042"$1"="$2"\042" }' ./ipv4settings.txt )
-bash: declare: `"ipv4_gateway=192.168.2.1"': not a valid identifier
-bash: declare: `"ipv4_routes=--"': not a valid identifier
-bash: declare: `"ipv4_route_metric=-1"': not a valid identifier
-bash: declare: `"ipv4_route_table=0': not a valid identifier
-bash: declare: `(unspec)"': not a valid identifier

...或者仅仅是价值部分：

$ awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print $1"=\042"$2"\042" }' ./ipv4settings.txt
ipv4_gateway="192.168.2.1"
ipv4_routes="--"
ipv4_route_metric="-1"
ipv4_route_table="0 (unspec)"
$ declare $( awk 'BEGIN{FS=":[[:space:]]+";OFS="="}{gsub("[.]|-","_",$1);$1=$1;print $1"=\042"$2"\042" }' ./ipv4settings.txt )
-bash: declare: `(unspec)"': not a valid identifier

如何处理declare带有空格的变量值？

bash
Martin Hope
vali hatami
Asked: 2024-12-23 07:50:15 +0800 CST
  • 5

我有一些 spring boot 微服务。我想使用 docker compose 运行它们。为了确保服务按顺序运行，我使用 .sh 文件在其他服务之前运行其中一个，因为其他微服务依赖于它。bash 脚本文件运行良好，但它会停止调用者容器。我在互联网上搜索并意识到必须在 bash 脚本文件末尾添加一些命令：/cnb/process/web或./cnb/lifecycle/launcher才能继续该过程。但显示此消息： /cnb/process/web: No such file or directory。

Spring Boot 版本：3.2.2

Docker 撰写：3.8

.sh 文件：

#!/bin/bash
# check-config-server-started.sh

apt-get update -y
yes | apt-get install curl

curlResult=$(curl -s -o -I -w "%{http_code}" http://cloud-config-server:8888/actuator/health)

echo "result status code:" "$curlResult"

while [[ ! $curlResult == "200" ]]; do
  >&2 echo "Config server is not up yet!"
  sleep 2
  curlResult=$(curl -s -o -I -w "%{http_code}" http://cloud-config-server:8888/actuator/health)
done
/cnb/process/web
#./cnb/lifecycle/launcher

Docker 撰写文件：

version: "3.8"

services:
  cloud-config-server:
    container_name: cloud-config-server-container
    build:
      dockerfile: ./cloud-config-server/Dockerfile
    ports:
      - "8888:8888"
    environment:
      - "SERVER_PORT=8888"
    networks:
      my-compose-net:
    restart: on-failure

  webapp:
    container_name: webapp-container
    build:
      dockerfile: ./webapp/Dockerfile
    ports:
      - "8081:8081"
    volumes:
      - "./check-config-server-started.sh:/usr/local/bin/check-config-server-started.sh"
    user: root
    entrypoint: [ "check-config-server-started.sh" ]
    environment:
      - "SPRING_CLOUD_CONFIG_URI=http://cloud-config-server:8888"
    depends_on:
      - cloud-config-server
    networks:
      my-compose-net:
    restart: on-failure

networks:
  my-compose-net:
bash