请问如何根据跨行条件截取文本信息？谢谢!

lxh3603

请问如何从文本信息获取满足 ptempdb_pages >= 10 的 spid?
例如，如下文本信息中，满足要求的是 spid 3 (ptempdb_pages = 10 )和spid 4 (ptempdb_pages = 20 )。

********  Active PSS Structures  ********
spid: 1, kpid: 1361641792, pss: 0x1002dc78e48, proc: 0x10034c22ea8
pmemusage: 3, ptempdb_pages: 3
  User: NMSuser, Command: 0 AWAITING COMMAND
  Status: (0x00010000 (P_USERPROC))
*****************************************
spid: 2, kpid: 327685, pss: 0x1002dc8e888, proc: 0x10034bb6050
pmemusage: 0, ptempdb_pages: 0
  Command: 0 DEADLOCK TUNE
  Status: (0x00000200 (P_SYSTEMPROC))
*****************************************
spid: 3, kpid: 393222, pss: 0x1002dca42c8, proc: 0x10034bb65d8
pmemusage: 0, ptempdb_pages: 10
  Command: 0 ASTC HANDLER
  Status: (0x00000200 (P_SYSTEMPROC))
*****************************************
spid: 4, kpid: 458759, pss: 0x1002dcb9d08, proc: 0x10034bb6b60
pmemusage: 0, ptempdb_pages: 20
  Command: 0 ASTC HANDLER
  Status: (0x00000200 (P_SYSTEMPROC))
*****************************************
spid: 5, kpid: 524296, pss: 0x1002dccf748, proc: 0x10034bb70e8
pmemusage: 0, ptempdb_pages: 0
  Command: 0 ASTC HANDLER
  Status: (0x00000200 (P_SYSTEMPROC))
*****************************************

personball

回复 1# lxh3603

1.     personball@vostro:SHELL$grep -oP "[\*]+|spid: \d+,|ptempdb_pages: \d+" file|awk -F'[:,]' 'BEGIN{RS="[*]+"} $4>=10{print $1":"$2}'
   
2. 
   
3. spid: 3
   
4. 
   
5. spid: 4
   
6. personball@vostro:SHELL$
   

复制代码

lxh3603

回复 2# personball


谢谢!

SOLARIS 10系统，grep没有 oP 参数。

bash-3.00# bash -version
GNU bash, version 3.00.16(1)-release (sparc-sun-solaris2.10)
Copyright (C) 2004 Free Software Foundation, Inc.

bash-3.00# grep -h
Usage: grep -hblcnsviw pattern file . . .

personball

回复 3# lxh3603

1. personball@vostro:SHELL$awk -F'[:, ]+' 'BEGIN{RS="[*]+"}$11>=10{print $1" "$2 }' file
   
2. 
   
3.         spid 3
   
4. 
   
5.         spid 4
   
6. personball@vostro:SHELL$
   

复制代码

lxh3603

回复 4# personball

谢谢!

难道最近俺RP出问题了？还是有问题：太长了。

awk: record `spid: 1, kpid: 13616...' too long
或
/usr/xpg4/bin/awk: line 0 (NR=1): Record too long (LIMIT: 19999 bytes)
   

lxh3603

回复 5# lxh3603


nawk只打出第一个满足条件的，不是所有的。

jiejie455

1. awk -vFS="[:, \t]+" '/spid/{spid=$2};/ptempdb_pages/&&$4>=10{print "spid: "spid}' data

复制代码

lxh3603

回复 7# jiejie455


谢谢! 这个在我的环境中测试通过了，没有问题。