shell: 替换文件中的部分行

Question

我是 shell 编程新手，必须完成以下任务。我在第 28 行（静态）有一个包含以下行的文件。

page.sysauth = "Admin"

我想在每次创建 sysauth 的新条目时使用 shell 脚本替换这一行。

page.sysauth = {"Admin", "root"} page.sysauth = {"Admin", "root", "newAdmin"} 等等

我还想从这个 sysauth 变量中删除条目

page.sysauth = {"Admin", "root", "newAdmin"}
page.sysauth = {"Admin", "root"}
page.sysauth = "Admin"

请指点如何实现。

编辑：感谢您的投入：假设：应该存在第一个条目。例如：page.sysauth="Admin" page.sysauth=______（空）时脚本失败。

这是我的工作脚本sysauth_adt.sh

#!/bin/bash
add () {
    sed -i~ -e '28 { s/= "\(.*\)"/= {""}/; # Add curlies to a single entry.
                     s/}/,"'"$entry"'"}/     # Add the new entry.
                   }' "$file"
}

remove () {
    sed -i~ -e '28 { s/"'"$entry"'"//;      # Remove the entry.
                     s/,}/}/;               # Remove the trailing comma (entry was last).
                     s/{,/{/;               # Remove the leading comma (entry was first).
                     s/,,/,/;               # Remove surplus comma (entry was inside).
                     s/{"\([^,]*\)"}/""/  # Remove curlies for single entry.
                   }' "$file"
}

if (( $# == 3 )) ; then
    file=
    action=
    entry=
    if [[ $action == add ]] ; then
        if head -n28  | tail -n1 | grep -q  ; then
                echo 0 
        else
                add
        fi
    elif [[ $action == remove ]] ; then
        if head -n28  | tail -n1 | grep -q  ; then
                remove
        else
                echo 0
        fi
    fi
else
    echo "Usage: ${0#*/} file (add | remove) entry" >&2
    exit 1
fi

Answer 1

如果您的条目总是没有逗号的单个单词，您可以使用简单的 sed 脚本：

#!/bin/bash
add () {
    sed -i~ -e '28 { s/= "\(.*\)"/= {""}/; # Add curlies to a single entry.
                     s/}/,"'"$entry"'"}/     # Add the new entry.
                   }' "$file"
}

remove () {
    sed -i~ -e '28 { s/"'"$entry"'"//;      # Remove the entry.
                     s/,}/}/;               # Remove the trailing comma (entry was last).
                     s/{,/{/;               # Remove the leading comma (entry was first).
                     s/,,/,/;               # Remove surplus comma (entry was inside).
                     s/{"\([^,]*\)"}/""/  # Remove curlies for single entry.
                   }' "$file"
}

if (( $# == 3 )) ; then
    file=
    action=
    entry=
    if [[ $action == add ]] ; then
        add
    elif [[ $action == remove ]] ; then
        remove
    fi
else
    echo "Usage: ${0#*/} file (add | remove) entry" >&2
    exit 1
fi

脚本在添加或删除条目时不检查条目是否已存在于列表中。对于更复杂的任务，我可能会切换到真正的编程语言。

Answer 2

awk -F '[[:blank:]]+=[[:blank:]]+' '
   # load every single entry
   NR != 28 && FNR == NR &&  ~ /^page.sysauth$/ && [=10=] !~ /\{.*\}/ { aSysAdd[ ++SysAdd] = }
   # load special line 28
   NR == 28 {
      gsub( /^{|}$/, "", Datas = )
      SysOrg = split( Datas, aSysOrg, /[[:blank:]]*,[[:blank:]]*/)
      }
   # print standard lines
   FNR != NR &&  !~ /^page.sysauth$/ { print }
   # print line 28 (new version)
   FNR != NR && FNR == 28 {
      printf "page.sysauth = {" aSysOrg[ 1]
      for( i = 2; i <= SysOrg; i++) printf ", %s", aSysOrg[ i]
      for( i = 1; i <= SysAdd; i++) printf ", %s", aSysAdd[ i]
      print "}"
   # don't print other page.sysauth
   }
 ' YourFile YourFile > YourFile.new
mv YourFile.new YourFile

使用 awk 进行 2 次读取（双 YourFile 不是错误且是必需的）

shell: 替换文件中的部分行

shell: Replacing a part of Line in a file

bash

shell

awk

sed

sh