Parse multiple variable from text file with bash

Goni · 07-13-2010, 01:52 AM

Hello everyone

I am trying to think of a logic where my file contains some data I had to read and do some processing. Issue is that file contains data multiple times. For example:

:::::::::::
var1=value1
var2=value2
:::::::::::
var1=value3
var2=value4
:::::::::::

I have to read first paragraph of variables and do some processing and then move on until the end of file. Variable names are same in whole file but for each paragraph the value is different. I can't think of a logic to attain this task. How can I do it? It should be a simple bash script, but I am not able to work out. Need help.

Thank you

jf.argentino · 07-13-2010, 02:05 AM

I'd "grep" for one variable name piped with "head" and "tail" with the "-n" argument, then you get the variable line from the paragraph you want. Then awk to parse the name in a couple of name / value

colucix · 07-13-2010, 02:25 AM

In bash an useful statement to parse a file line-by-line is the while read loop. You can read two variables per line, using the equal sign as Input Field Separator. Inside the loop you can do whatever you want with the obtained variables, e.g.

Code:

#!/bin/bash
while IFS="=" read name value
do
  echo name is $name
  echo value is $value
done < file

Anyway, you may want to exclude some lines from the input (for example those ones with colons in your example). In that case you can feed the loop using process substitution. Just think at a command to retrieve only the lines you want to read, e.g.

Code:

done < <(grep -v ':' file)

Hope this helps.