Shell script: Merge rows.

Author: gnuhpc
Source: http://www.cnblogs.com/gnuhpc/

There is a text:
2006
China
Sichuan
042834 1 2 3
042835 4 5 6
042836 7 8 9
2007
China
Chongqing
042837 1 2 3
042838 4 5 6
042839 7 8 9
......
......
To merge
042834 1 2 3 2006 Sichuan, China
042835 4 5 6 2006 Sichuan, China
042836 2006 Sichuan, China
042837 1 2 3 2007 Chongqing, China
042838 4 5 6 2007 Chongqing, China
042839 2007 Chongqing, China
The script is as follows:
#! /Bin/bash
######################################## #################################
# Author: pchuang@cn.ibm.com
# Created time: Sun 05 Apr 2009 10:04:51 am CST
# File name: process. Sh
# Description: a script for the processing of the text
######################################## #################################
File = $1
I = 1
Lines [0] = ""
If [! -S $1]; then
Echo "Please specify a valid text file"
Exit 1
Fi
While read line
Do
Lines [$ I] = "$ line"
If [$ I-EQ 6]; then
Lines [$ I] = "$ line"
For J in $ (SEQ 4 6); do
Echo $ {lines [$ J] }$ {lines [1] }$ {lines [2] }$ {lines [3] }>> result.txt
Done
Let I = 0
Fi
Let I = $ I + 1
Done <$1
Run:
$./Process. Sh text
$ More result.txt
042834 1 2 3 2006 Sichuan, China
042835 4 5 6 2006 Sichuan, China
042836 2006 Sichuan, China
042837 1 2 3 2007 Chongqing, China
042838 4 5 6 2007 Chongqing, China
042839 2007 Chongqing, China
Later, a cool man used awk to complete the process:
Awk 'in in {I = 0} {If (NF! = 4) {H [I ++] = $0; next} else {I = 0; print $0 "" H [0] "H [1]" H [2]}'

Author: gnuhpc
Source: http://www.cnblogs.com/gnuhpc/