use utf8;
use Encode;
use IO::Handle;
use File::Slurp;
STDOUT->autoflush(1);

my $src = encode('gbk', "原始文本.txt");
my @source = read_file( $src );
my $idx = 0;

#去掉抬头
do { shift @source } until ( $source[0]=~/======/  );
shift @source;

my $flag = 0;
while ( $idx <= $#source )
{
    ($DNI, $Material) = (substr($source[$idx], 9, 3), substr($source[$idx], 76, 20));
    $Material=~s/\s//g;

    if ( $DNI eq "DNI" )
    {
        print $Material,"\n";
        $flag = 1;
    }
    else
    {
        if ( $DNI eq "   " and $flag == 1 and $Material ne "" )
        {
            print $Material,"\n";
        }
        else
        {
            $flag = 0;
        }
    }

    $idx++;
}
复制代码

@echo off&setlocal enabledelayedexpansion

set n=2
for /f "tokens=1,2,3,4,5,6 delims= " %%a in (src.txt) do (
   set /a n+=1
   if not "x%%f"=="x" ( if "x%%b"=="x===" (
      call :headline "%%a" "%%b" "%%c" "%%d" "%%e" "%%f"
      goto :end_head
   ))
)

:end_head
echo COLUMN_LEN:%id_len%, %type_len%, %num_len%, %name_len%, %qty_len%, %des_len%
echo.

set is_data_end=0
for /f "skip=%n% delims=" %%a in (src.txt) do (
   call :AnalysisLine "%%a"
   if %is_data_end% NEQ 0 goto :end_data
)

:end_data
pause
goto :EOF

:AnalysisLine
set "AL_line=%~1"
set AL_id=!AL_line:~0,%id_len%!
set AL_type=!AL_line:~%type_start%,%type_len%!
set AL_des=!AL_line:~%des_start%,%des_len%!

if "x%AL_id%"=="x TOTAL" (
  set is_data_end=1
  goto :EOF
)

if "x%AL_type%"=="x   " (
  set "AL_type=%AL_type_LAST%"
)

set AL_type_LAST=%AL_type%

if "x%AL_type%"=="xDNI" (
  echo [%AL_des%]
)
goto :EOF

:headline
call :Len %1
set id_len=%Len_Ret%
call :Len %2
set /a type_start=%id_len%+3
set type_len=%Len_Ret%
call :Len %3
set num_len=%Len_Ret%
call :Len %4
set name_len=%Len_Ret%
call :Len %5
set qty_len=%Len_Ret%
call :Len %6
set /a des_start=%id_len%+%type_len%+%num_len%+%name_len%+%qty_len%+5*3
set des_len=%Len_Ret%
goto :EOF


:Len
  set LenVar_i=0
  set LenVar_Str=%~1
  if "x%LenVar_Str%"=="x" (
    set /a Len_Ret=0
    goto :EOF
  )
  :Len_Loop
    set LenVar_c=!LenVar_Str:~%LenVar_i%,1!
    if "x%LenVar_c%"=="x" goto :Len_EndLoop
    set /a LenVar_i+=1
    goto :Len_Loop

  :Len_EndLoop
  set /a Len_Ret=LenVar_i
  goto :EOF
复制代码

lines = File.readlines('src.txt')
dataline = false
last_type = ''
des_start = des_len = 0

lines.each do |line|
  line.chomp! # 去掉结尾Windows上可能多的\r回车符
  if dataline == false && line[0, 6] == ' S.No.'
    line =~ /^( S.No.   DNI   .+Qty   )(Ref Des +)   COMPANY/       #正则匹配 标题
    des_start, des_len = $1.length, $2.length                       #保存Des的标题的开始和长度
    dataline = true
  elsif dataline == true
    type = line[9, 3]    #DNI列固定?
  end
  break if line[0, 6] == ' TOTAL'     #遇到TOTAL结束
  type = last_type if type == '   '   #如果type是3个空格，那么用上次type
  last_type = type                    #设置当前type为最后type
  puts "[#{line[des_start, des_len]}]" if type == 'DNI'
end
复制代码

@echo off
setlocal enabledelayedexpansion
set "file=原始文本.txt"
set /a flag = 0

for /f "tokens=* delims=" %%a in ('type %file% ^| more +7 ') do (
	set "str=%%a"
	set DNI=!str:~9,3!
	set Material=!str:~76,20!
	set Material=!Material: =!
	if "!DNI!" == "DNI" (
		echo !DNI! !Material!
		set /a flag = 1
	) else (
		if "!DNI!!flag!" == "   1" (
			if not "!Material!" == "" (
				echo !DNI! !Material!
			)
		) else (
			set /a flag=0
		)
	)
)
pause
复制代码