一个很不错的ｆｔｐ脚本，但是我不会改哦

liuxuexin_xxlhz

1. #!/bin/sh
   
2. # ftpget.sh
   
3. 
   
4. 
   
5. 
   
6. SPSC_SERVER=192.168.0.1
   
7. FTP_USERNAME=11
   
8. FTP_PASSWORD=11
   
9. FILE_PRENAME=mobile
   
10. 
    
11. Log ()
    
12. {
    
13.         echo [`date '+%Y%m%d %H:%M:%S'` $$] $*
    
14. }
    
15. 
    
16. if [ "-?" = "$1" ]
    
17. then echo usage: ftpgetsample.sh [-d dest_path]
    
18.         echo         dest_path: The local path in which files are stored, the recv path is default.
    
19.         echo    A sample: ftpget.sh -d /chroot/gateway
    
20. else
    
21.         exec>>/chroot/gateway/log/recv`date '+%Y%m%d'`.log 2>&1
    
22.         if [ "-d" = "$1" ]
    
23.                 then DEST_PATH=$2
    
24.                 else DEST_PATH=/chroot/gateway/
    
25.         fi
    
26. 
    
27.         cd $DEST_PATH
    
28.         Log "Begin to get filelist..."
    
29.        
    
30.         # You can NOT use mget and mdelete commands of ftp to handle this issue,
    
31.         # because when you begin to execute the mdelete command,
    
32.         # it's possible that a new file have been written into the disk,
    
33.         # just before the mdelete command start and after mget command had been finished.
    
34.         # as a result, firstly, you should get the file list,
    
35.         # secondly, get and delete each file in the list
    
36.        
    
37.         # Step 1: Get the file list from FTP server
    
38.         ftp -i -n $SPSC_SERVER <<MACRO_END1
    
39.           user $FTP_USERNAME $FTP_PASSWORD
    
40.           bin
    
41.           ls *$FILE_PRENAME.*  filelist
    
42.           bye
    
43. MACRO_END1
    
44. 
    
45.         if [ ! -s filelist ];then
    
46.            Log "No file received......"
    
47.            rm -f filelist
    
48.            exit
    
49.         fi
    
50.         # Step 2: Get and delete each file in circle.
    
51.         # "cut -b56-" means get the filename from 56th bytes in each line in filelist file,
    
52.         # you can adjust the number by yourself after read the filelist file.
    
53.         # In our testing environment of Turbo Linux and Red Hat Linux, the number is set to 56
    
54.         # In our testing environment of Solaris and AIX, the number is set to 1
    
55.         Log "transfering start..."
    
56.         for _FILENAME in `more filelist|cut -b56-` ; do
    
57.                 ftp -i -n $SPSC_SERVER <<MACRO_END2
    
58.                 user $FTP_USERNAME $FTP_PASSWORD
    
59.                 bin
    
60.                 get $_FILENAME $_FILENAME.tmp
    
61.                 delete $_FILENAME
    
62.                 bye
    
63. MACRO_END2
    
64.         mv $_FILENAME.tmp $_FILENAME
    
65.         Log "get $_FILENAME ok"  
    
66.         gunzip $_FILENAME
    
67.         Log "gunzip $_FILENAME ok"
    
68.         . /chroot/gateway/bin/subsync.sh
    
69.         done
    
70.         rm -f filelist
    
71.        
    
72.         Log "transfering completed."
    
73.        
    
74.         # Step 3: Handle files based on your business logic
    
75.         # The following handle program should be written by yourself !!!
    
76.         #for _FILENAME in `more filelist|cut -b56-` ; do
    
77.         #        java HandleUnsubRecord ......
    
78.         #done
    
79. fi
    

复制代码

我想从服务器上把文件取下来，然后将文件名写到一个.txt文件中，为了下次再取的时候可以比较一下服务器上的文件我是否取过了，同时在我取完就不删除服务器上的文件了，有没有熟悉方面的人指点我一下，我是ｓｈｅｌｌ新人。
１，取来的文件名写到.txt文件中
２，比较服务器上的文件和我.txt中的文件
３，取消删除操作（这个我还会一点，呵呵）

ZealeS

用ftp执行
ls *.txt nowfile
就会把ftp的目录下的txt文件名打印到本地目录下的nowfile文件。

un_xxx

原帖由 "liuxuexin_xxlhz" 发表：
#!/bin/sh
# ftpget.sh



SPSC_SERVER=192.168.0.1
FTP_USERNAME=11
FTP_PASSWORD=11
FILE_PRENAME=mobile

Log ()
{
   echo [`date '+%Y%m%d %H:%M:%S'` $$] $*
}

if [ "-?" = "$1" ]
then echo usage: ftpgetsample.sh [-d dest_path]
   echo    dest_path: The local path in which files are stored, the recv path is default.
   echo    A sample: ftpget.sh -d /chroot/gateway
else
   exec>>/chroot/gateway/log/recv`date '+%Y%m%d'`.log 2>&1
   if [ "-d" = "$1" ]
      then DEST_PATH=$2
      else DEST_PATH=/chroot/gateway/
   fi

   cd $DEST_PATH
   Log "Begin to get filelist..."
   
   # You can NOT use mget and mdelete commands of ftp to handle this issue,
   # because when you begin to execute the mdelete command,
   # it's possible that a new file have been written into the disk,
   # just before the mdelete command start and after mget command had been finished.
   # as a result, firstly, you should get the file list,
   # secondly, get and delete each file in the list
   
   # Step 1: Get the file list from FTP server
        ftp -i -n $SPSC_SERVER <<MACRO_END1
          user $FTP_USERNAME $FTP_PASSWORD
          bin

1. ls *$FILE_PRENAME.*  filelist.$$.txt #改成txt文件

复制代码

原帖由 "liuxuexin_xxlhz" 发表：
           bye
MACRO_END1

   if [ ! -s filelist ];then
      Log "No file received......"
      rm -f filelist
      exit
   fi

1. comm -1 filelist.$$.txt filelist.txt  #比较新文件filelist.$$.txt与老文件filelist.txt

复制代码

原帖由 "liuxuexin_xxlhz" 发表：
#   # Step 2: Get and delete each file in circle.
   # "cut -b56-" means get the filename from 56th bytes in each line in filelist file,
   # you can adjust the number by yourself after read the filelist file.
   # In our testing environment of Turbo Linux and Red Hat Linux, the number is set to 56
   # In our testing environment of Solaris and AIX, the number is set to 1
   Log "transfering start..."
   for _FILENAME in `more filelist|cut -b56-` ; do
                ftp -i -n $SPSC_SERVER <<MACRO_END2
                user $FTP_USERNAME $FTP_PASSWORD
                bin
                get $_FILENAME $_FILENAME.tmp

1. #           delete $_FILENAME  不要删除文件

复制代码

原帖由 "liuxuexin_xxlhz" 发表：
                bye
MACRO_END2
   mv $_FILENAME.tmp $_FILENAME
   Log "get $_FILENAME ok"
   gunzip $_FILENAME
   Log "gunzip $_FILENAME ok"
   . /chroot/gateway/bin/subsync.sh
   done
   rm -f filelist
   
   Log "transfering completed."

1. mv filelist.$$.txt filelist.txt

复制代码

原帖由 "liuxuexin_xxlhz" 发表：
# Step 3: Handle files based on your business logic
   # The following handle program should be written by yourself !!!
   #for _FILENAME in `more filelist|cut -b56-` ; do
   #   java HandleUnsubRecord ......
   #done
fi  

liuxuexin_xxlhz

怎么样给comm -1 filelist.$$.txt filelist.txt  #比较新文件filelist.$$.txt与老文件filelist.txt加一个ｉｆ语句来判断我的ｆｉｌｅｌｉｓｔ.txt中如果有这个文件了，就不再ｇｅｔ了

liuxuexin_xxlhz

ｑｕｅｓｔｉｏｎ：
１，

使用comm的前提是,两个文件必须是经过排序过的
基本格式:
comm -1|-2|-3 file1 file2
-1:不显示在file1中存在的行
-2:不显示在file2中存在的行
-3:不显示在file1和file2中都存在的行
例如不显示file1,file2都有的行
comm -3 file1 file2

请问comm有没有返回值，我想加一个ｉｆ来判断下面是否执行ｇｅｔ
２，

mv filelist.$$.txt filelist.txt

如果这样的话我每次ｇｅｔ下来的文件内容就重新作为filelist.txt的内容了，能不能达到filelist.txt中的内容一步一步的增多，作为前面多次ｇｅｔ的内容，以便我来比较服务器上的文件我曾经是否取过了

un_xxx

1. diff filelist.$$.txt filelist.txt |grep "<" |awk '{print $2}' >file to get
   
2. rm -f filelist.$$.txt
   
3. # # Step 2: Get and delete each file in circle.
   
4. # "cut -b56-" means get the filename from 56th bytes in each line in filelist file,
   
5. # you can adjust the number by yourself after read the filelist file.
   
6. # In our testing environment of Turbo Linux and Red Hat Linux, the number is set to 56
   
7. # In our testing environment of Solaris and AIX, the number is set to 1
   
8. Log "transfering start..."
   
9. for _FILENAME in `more filetoget|cut -b56-` ; do
   
10. ftp -i -n $SPSC_SERVER <<MACRO_END2
    
11. user $FTP_USERNAME $FTP_PASSWORD
    
12. bin
    
13. get $_FILENAME $_FILENAME.tmp
    
14. #           delete $_FILENAME  不要删除文件
    
15. bye
    
16. MACRO_END2
    
17. mv $_FILENAME.tmp $_FILENAME
    
18. Log "get $_FILENAME ok"
    
19. gunzip $_FILENAME
    
20. Log "gunzip $_FILENAME ok"
    
21. . /chroot/gateway/bin/subsync.sh
    
22. done
    
23. cat filetoget >>filelist.txt
    
24. rm -f filetoget
    
25. 
    
26. Log "transfering completed."
    
27. # Step 3: Handle files based on your business logic
    
28. # The following handle program should be written by yourself !!!
    
29. #for _FILENAME in `more filelist|cut -b56-` ; do
    
30. # java HandleUnsubRecord ......
    
31. #done
    
32. fi

复制代码

un_xxx

从filelist.txt比较开始的前边的都没有修改直接粘贴上就可以了。

liuxuexin_xxlhz

能不能解释一下

diff filelist.$$.txt filelist.txt |grep "<" |awk '{print $2}' >file to get

中的　awk '{print $2}' ，当我执行完sh ftpget.sh
之后我的filelist.txt　文件中都是

-rw-rw-rw-
-rw-rw-rw-
-rw-rw-rw-
-rw-rw-rw-

我弄了半天这个ｄｉｆｆ，可就是区分不出不同的文件来。
我理解你的意思是用ｄｉｆｆ之后把filelist.$$.txt和 filelist.txt中的内容进行比较之后，选出”<“的文件重定向到　filetoget，但是定向的结果有问题，我把$2换成$３之后，我的filelist.txt文件的内容就是

1
1
1
1

真的不知道是为什么了，我不会用ａｗｋ，请指教一下了

un_xxx

不要意思，我当时没有考虑filelist的内容，当在FTP的时候生成的filelist.$$.txt的格式如下：

1. -rw-r--r--    1 0        0            1538 Sep 29 08:12 GG.jpg
   
2. -rw-r--r--    1 0        0           26307 Dec 22 02:38 Img223132522.jpg
   
3. drwxr-xr-x    3 0        0            4096 Dec 22 02:41 Projects
   
4. -rw-r--r--    1 0        0            1617 Sep 07 07:06 Untitled1.CRASHED
   
5. -rw-r--r--    1 0        0            1264 Jul 31 18:42 anaconda-ks.cfg
   
6. -rw-r--r--    1 0        0           11919 Nov 24 00:46 asp2php-gtk-0.76.2-5.i386.rpm
   
7. -rw-r--r--    1 0        0           28178 Nov 24 00:46 authconfig-gtk-4.3.4-1.i386.rpm
   
8. -rw-r--r--    1 0        0           16948 Nov 12 05:31 autorun

复制代码

不好意思，我再修改下。首先提取filelist.$$.txt只提取最后一项文件名

1. awk '{print $9}' filelist.$$.txt >filelist.name.txt

复制代码

需要说明的是filelist.txt必须也是只包含文件名的文件，这样有利于diff进行比较。

1. diff filelist.name.txt filelist.txt |grep "<" |awk '{print $2}' >filetoget

复制代码

注意把产生的临时文件（filelist.name.txt和filelist.$$.txt）删除这样就把filelist.name.txt文件中独有的文件名字复制到filetoget这个文件中。这样我想应该能行了吧，嘿嘿

liuxuexin_xxlhz

不好意思，昨天病了，今天继续研究这个东东！