（求助）使用shell命令按標(biāo)識分割文件

Pugna_zfr

一個文件內(nèi)容如下：
Service "syncdb" has 1 instance(s).
  Instance "syncdb", status READY, has 1 handler(s) for this service...
Service "syncdbXDB" has 2 instance(s).
  Instance "syncdb1", status READY, has 1 handler(s) for this service...
  Instance "syncdb2", status READY, has 1 handler(s) for this service...
Service "syncdb_XPT" has 1 instance(s).
  Instance "syncdb", status READY, has 1 handler(s) for this service...



使用什么shell命令可以將上面的文件分割為以下3個文件：

第一個文件：
Service "syncdb" has 1 instance(s).
  Instance "syncdb", status READY, has 1 handler(s) for this service...
第二個文件：
Service "syncdbXDB" has 2 instance(s).
  Instance "syncdb1", status READY, has 1 handler(s) for this service...
  Instance "syncdb2", status READY, has 1 handler(s) for this service...
第三個文件：
Service "syncdb_XPT" has 1 instance(s).
  Instance "syncdb", status READY, has 1 handler(s) for this service...

也就是按Service開頭為分割標(biāo)識，相鄰的兩個Service之間的行數(shù)是不確定的，可能是1行，也可能是n行。

謝謝！

polyahu

1. awk '{if($1 == "Service" && $2 == "\"syncdb\""){n=$4+1;}else{;};if(n>0){print $0;n=n-1}}' data.txt > output
   
2. awk '{if($1 == "Service" && $2 == "\"syncdbXDB\""){n=$4+1;}else{;};if(n>0){print $0;n=n-1}}' data.txt > output.XDB
   
3. awk '{if($1 == "Service" && $2 == "\"syncdb_XPT\""){n=$4+1;}else{;};if(n>0){print $0;n=n-1}}' data.txt > output._XPT

復(fù)制代碼

僅供參考。

laliheyi

回復(fù) 1# Pugna_zfr


    awk '/^Service/{file=$2;gsub(/\"/,"",file)}{print $0 > file}' service.txt

Pugna_zfr

回復(fù) 2# polyahu


    謝謝！
    我用 awk '{if($1 == "Service" && $2 !~/XDB/ && $2 !~/_XPT/ && $2 !~/PLSExtProc/){n=$4+1;}else{;};if(n>0){print $0;n=n-1}}' inputfile 基本達到要求了，就是有個點遺憾，不能自動按Service名稱分割。不過還是很不錯了！3Q

Pugna_zfr

回復(fù) 3# laliheyi


    試了下，好像有點問題。
    -bash-3.2$ lsnrctl status | awk '/^Service/{file=$2;gsub(/\"/,"",file)}{print $0 > file}'
    awk: (FILENAME=- FNR=1) fatal: expression for `>' redirection has null string value

    -bash-3.2$ awk --version
    GNU Awk 3.1.5

Pugna_zfr

回復(fù) 4# Pugna_zfr


    哦，不好意思，我剛才搞錯了。我把lsnrctl status的輸出內(nèi)容格式化成1L中的那樣是正常的。
    謝謝！ :wink:

laliheyi

回復(fù) 5# Pugna_zfr


   {:2_166:}  lsnrctl status 結(jié)果前面那部分結(jié)果你的需求中沒有給出。如果那部分不需要可以如此
awk '/^Service/{file=$2;gsub(/\"/,"",file)}file!=""{print $0 > file}'

WilliBhamlll

回復(fù) 1# Pugna_zfr

1. $awk '/^Service/{++i}{print >"file"i}' file
   
2. 
   
3. $cat file
   
4. Service "syncdb" has 1 instance(s).
   
5.   Instance "syncdb", status READY, has 1 handler(s) for this service...
   
6. Service "syncdbXDB" has 2 instance(s).
   
7.   Instance "syncdb1", status READY, has 1 handler(s) for this service...
   
8.   Instance "syncdb2", status READY, has 1 handler(s) for this service...
   
9. Service "syncdb_XPT" has 1 instance(s).
   
10.   Instance "syncdb", status READY, has 1 handler(s) for this service...
    
11. 
    
12. $cat file1
    
13. Service "syncdb" has 1 instance(s).
    
14.   Instance "syncdb", status READY, has 1 handler(s) for this service...
    
15. $cat file2
    
16. Service "syncdbXDB" has 2 instance(s).
    
17.   Instance "syncdb1", status READY, has 1 handler(s) for this service...
    
18.   Instance "syncdb2", status READY, has 1 handler(s) for this service...
    
19. $cat file3
    
20. Service "syncdb_XPT" has 1 instance(s).
    
21.   Instance "syncdb", status READY, has 1 handler(s) for this service...

復(fù)制代碼

Pugna_zfr

回復(fù) 7# laliheyi


    嗯。

命令中的 print $0 > file 看不太明白。file是前面的service名稱，$0是匹配到以Service開頭的每一行嗎？
我改了下這個命令，感覺$0是把每一行都匹配了，不管是否以Service開頭。但是，如果$0是每一行的話，直接用 > 到文件，居然沒有把原文件的內(nèi)容（即剛剛輸出到文件的上一行內(nèi)容）覆蓋？？？
a.txt中是格式化好的文本內(nèi)容。

-bash-3.2$ awk '/^Service/{file=$2;gsub(/\"/,"",file)}{print "test line\n"$0}' a.txt
test line
Service "PLSExtProc" has 1 instance(s).
test line
  Instance "PLSExtProc", status UNKNOWN, has 1 handler(s) for this service...
test line
Service "syncdb" has 1 instance(s).
test line
  Instance "syncdb", status READY, has 1 handler(s) for this service...
test line
Service "syncdbXDB" has 1 instance(s).
test line
  Instance "syncdb", status READY, has 1 handler(s) for this service...
test line
Service "syncdb_XPT" has 1 instance(s).
test line
  Instance "syncdb", status READY, has 1 handler(s) for this service...

Pugna_zfr

回復(fù) 8# WilliBhamlll


    這個 awk '/^Service/{++i}{print >"file"i}' file 是不是匹配以Service開頭的內(nèi)容，如果匹配到就執(zhí)行{++i}{print >"file"i} ，直到遇到下一個匹配Service開頭的行，再執(zhí)行{++i}{print >"file"i} ，然后一直這么循環(huán)下去，直到讀取完文件的內(nèi)容 ？？