regex - 脚本在空格上失败，我应该编写 sed 来转义路径或更改脚本以在目录上工作

Question

我有一个 samba 目录共享(~1500)示例:101284_2012110634 Sachine 130 AB i HeisMakeBoravia(橄榄色)202373_2012110640 Snowchine 7 AB i 法兰德斯300738_2012110658_Machine VP 7 AB i 重命名 v 500185

我正在编写一个脚本来解析目录并从目录名称中获取一些数据(序列号、名称、注释)，并枚举一些文件，特别是列表中目录的子目录(始终相同)。我通过 ls > results.txt 生成了目录列表 我通过 parse.sh results.txt

运行脚本

我完成了脚本的前半部分，它从目录名称中获取我需要的数据(省略了一些行，因为它们不相关)并且它有效。我对脚本的第二部分有问题，它首先查找子目录是否存在，然后查找 file 类型的文件并获取它们的名称和 md5 哈希值。如果在自己没有空格的目录中运行，第二部分将独立工作。

#!/bin/bash
# manually set working directory /sample or /rlisti must combine with parse.sh sample.txt or results.txt
WORKING_DIR='/sample'

# read from input file which is generated from ls dir > results.txt
# usage ./parse.sh results.txt

while IFS= read line || [[ -n $line ]]; do

  # some tests which find variables from $line

  MACHINE_SN=$(echo "$line" | sed 's/\([0-9][0-9][0-9][0-9][0-9][0-9]\).*/\1/')
  BOARD_SN=$(echo "$line" | sed 's/^[0-9][0-9]*[ _]*\([0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]\).*/\1/')

  if echo $line | grep -iqF air
  then
    AIR='1'
  else
    AIR='0'
  fi

  # tests to find and hash particular files in sub dirs
  # test if there is directory Actual_Program_and_Database in current line.

  if [ -d "$WORKING_DIR"/"$line"/Actual_Program_and_Database ]
  then

    # if dir exists set SW_DIR to 'Actual_Program_and_Database

    SW_DIR='Actual_Program_and_Database'

    # count files in dir, there should be 3 or 4, good to know if there are less or more as those are exceptions

    NUM_FILES=$(ls "$WORKING_DIR"/"$line"/Actual_Program_and_Database | wc -l)
    DB_COUNTER=1 # count Clarion DB files in dir

    # test each file - this works on it's own when run in same directory with files
    # but fails if lines contain spaces

    for FILENAME in "$WORKING_DIR"/"$line"/Actual_Program_and_Database/*
    do

      # bellow works when script is run from same directory
      # FILENAME is just file without it's leading path - so I'm trying to merge it with line

      FILENAME_PATH="$WORKING_DIR"/"$line"/Actual_Program_and_Database/"$FILENAME"

      # files of type Clarion are DB, there is DB1 and DB2 if there is also DB3 that is exception

      if file "$FILENAME" | grep -q Clarion
      then

        # Create variables with name DB_PATH1 DB_PATH2 etc
        # eval works on it's own but breaks if paths contain spaces or ()&

        eval 'DB_PATH'$DB_COUNTER="$FILENAME_PATH" # path to filename for use with md5sum
        eval 'DB_NAME'$DB_COUNTER=$(basename "$FILENAME_PATH") # name of DBx file
        eval 'DB_HASH'$DB_COUNTER=$(md5sum "$FILENAME_PATH" | awk '{ print $1 }') # create hash  of DBx
        ((DB_COUNTER ++)) # increase DB counter

      # files of type POSIX are SCRIPTs there should be only one

      elif file "$FILENAME" | grep -q POSIX
      then
        SCRIPT_PATH=$FILENAME_PATH
        SCRIPT_NAME=$(basename $FILENAME_PATH)
        SCRIPT_HASH=$(md5sum $FILENAME_PATH | awk '{ print $1 }')

      # files of type CRLF are LISTs there should be only one

      elif file "$FILENAME" | grep -q CRLF
      then
        LIST_PATH=$FILENAME_PATH
        LIST_NAME=$(basename $FILENAME_PATH)
        LIST_HASH=$(md5sum $FILENAME_PATH | awk '{ print $1 }')
      else
        UNKNOWN_FILE='1'
      fi
    done

  # if there isn't directory Actual_Program_and_Database in current line set dir to 0
  # there are not enough exceptions to program for them

  else
    SW_DIR='0'
  fi

  # Print results (for now - will write SQL statements to populate database once script works)

  echo 'serial='$SERIAL
  echo 'machine sn =   '$MACHINE_SN # Valmar machine SN
  echo 'board sn =     '$BOARD_SN # electronic board SN
  echo 'NUM_FILES =    '$NUM_FILES # number of files in directory
  echo 'NUM_DBS =      '$NUM_DBS # number of files in directory - not working
  echo 'DB_COUNTER=  '$DB_COUNTER
  echo 'SCRIPT_NAME= '$SCRIPT_NAME
  echo 'SCRIPT_HASH= '$SCRIPT_HASH
  echo 'LIST_NAME=   '$LIST_NAME
  echo 'LIST_HASH=   '$LIST_HASH
  echo 'DB_NAME1=    '$DB_NAME1
  echo 'DB_HASH1=    '$DB_HASH1
  echo 'DB_NAME2=    '$DB_NAME2
  echo 'DB_HASH2=    '$DB_HASH2
  echo 'UNKNOWN_FILE='$UNKNOWN_FILE

  # added sleep so I can stop script when it fails to see results

  sleep 10

done < "$1"

我是否应该将 sed 添加到开始创建 LINE_ESCAPED ，这将转义行中的所有空格括号并在第二部分中使用它？或者我应该重写脚本以在共享上工作而不是在 results.txt 上工作，这甚至可以工作吗？

Answer 1

我的引用有一些错误，我正在附加路径。我之前使用eval是因为我需要db1= db2=并且我想使用eval来生成变量数字。我从 eval 切换到 array[counter]，因为我避免了 eval，而 array 满足了我的需要。不了解数组 - 这是我的第一个更严肃的 bash 脚本。

我还修复了所有大写变量，并使用建议仅测试文件 $filename 一次。这是我现在可以使用的代码。谢谢大家的建议。

if [ -d "$working_dir"/"$line"/Actual_Program_and_Database ]
then
  sw_dir='Actual_Program_and_Database'
  num_files=$(ls "$working_dir"/"$line"/Actual_Program_and_Database | wc -l) #count files in dir
  db_counter=0 # count Clarion DB files in dir
  for filename in "$working_dir"/"$line"/Actual_Program_and_Database/*
  do
  file_type=$(file "$filename")
  if [[ $file_type = *Clarion* ]]
    then
      db_path[$db_counter]=$filename
      db_name[$db_counter]=$(basename "$filename")
      db_hash[$db_counter]=$(md5sum "$filename" | awk '{ print $1 }')
      ((db_counter ++))
     elif [[ $file_type = *POSIX* ]]
     then
       script_path=$filename
       script_name=$(basename "$filename")
       script_hash=$(md5sum "$filename" | awk '{ print $1 }')
     elif [[ $file_type = *CRLF* ]]
     then
       list_path=$filename
       list_name=$(basename "$filename")
       list_hash=$(md5sum "$filename" | awk '{ print $1 }')
     else
       unknown_file='1'
  fi
  done
else
  sw_dir='0'
fi