如何从自定义行开始循环遍历bash中的文件？

Question

I have a file for which I want to loop through just a specific part of using a bash script. Here is an example of it's format:

FILE 1

(exclude)
. 

(Start loop from here)
.
.
.
(stop)

How do I loop through the file starting at the line I need and stop where required ? I assume I will need a for loop for this situation ? I do know the pattern for the first line if grep is required. Or the line number which is probably constant for all of my files (say starts at 5, stops at 9)

Answer 1

Set up a state machine using AWK that toggles processing of the lines ON when the start pattern is found and either

1. toggles processing OFF when the stop pattern is found (this allows processing more than one start/stop group

or

2. exits the script when the stop pattern is found (this terminates all processing and speeds things up when you don't care about anything after the stop pattern)

awk '/stop/  {process = 0}
     process {print $1, $5}
     /start/ {process = 1}' inputfile

This could be put on one line, but I've broken it apart here for easier reading.

process is being used as a boolean variable. If it's true, print a couple of fields from the current line.

/start/ and /stop/ represent regular expressions to match strings that you will use to delimit the section(s) of the file you wish to process. You could instead use line numbers, for example you could change the /start/ test to NR == 120 to begin processing on the 120th line.

If you want the beginning and ending lines to be included in the processing, just move the process line after the /start/ line and the /stop/ line to be the last (essentially, reverse their order).

As I've shown it, this script will process multiple start/stop blocks within the file. If you want to only process the first (or only) you can change the /stop/ line to:

/stop/  {exit)

Answer 2

Since you said that you have the line numbers, you can use tail . For instance, to loop through the file starting at line 42, you could do a

tail -n +42 YOUR_FILE | while line
do
    process_line "$line"
done

Answer 3

Example: extract user name and home directory from /etc/passwd starting at mail and ending at nobody .

cat /etc/passwd |
sed -n '/^mail/,/^nobody/p' |
while IFS=: read user pw uid gid geco home sh line; do
  echo "$user: $home"
done