Can someone help me with this shell scripting prob

**Fox** · 03-31-2002, 06:50 AM

Hi, I'm fairly new at shell scripting and I'm trying to create a small menu system for someone. It's a small script that is designed to analyse a file containing some text. (using Linux, Mandrake 7.2 to be precise).

I'm currently stuck on the following items:

1) I'm trying to get a specific letter count. For example, if someone wanted to count the number of 'f' in a text file, how would I do that.

2) How do I do aword count that can be specified, for example 'milk' in a text file.

3) A sentence count.

4) Add an extra line to the text

5) Replace a word in a text

I know it has something to do with grep, sed and wc, but I can't get around to piping them correctly.

Just one other thing, How do I get a menu system to work? I currently have a menu system that you can chose options such as:

[1] Open file
[2] List File
[3] Word count

I did this using functions but after the menu function has been executed once, then I call it from another function it doesn't seem to work, this is how my structure is:

function menu
{
menu code
option1 #this line calls the function option
}

menu #this line calls the function menu above

function option1
{
Option code

menu # this line is the problem, it calls the function, but the menu function doesn't work properly

}

Thank you for any help

**pam** · 04-05-2002, 08:25 PM

I don't know if you're still looking for answers to this, but I thought I'd have a bash anyway (no pun intended :

).

To count the f's in myfile.txt

Code:

sed -e 's/f/f\
/g' myfile.txt &#124;grep --count f

grep counts only lines, so the sed replaces all f's with f followed by newline (the new line in that sequence is quite deliberate). I guess if you research a bit, you could probably come up with something less clumsy

Similarly:

Code:

sed -e 's/\bmilk\b/\0\
/g' myfile.txt &#124;grep --count &quot;\bmilk\b&quot;

A sentence count might be (probably some flaws in my thinking here but what the hell)

Code:

sed -e 's/[\.\?\!]+/\0\
/g' myfile.txt &#124;grep --count &quot;[\.\?\!]&quot;

To assign the output from a command such as those above to a variable in your script, you need to surround the sequence with the backtick (`) character.

To add an extra line you could use:

Code:

cat &gt;&gt;myfile.txt

To replace a word, modify that sed above (the one that replaces milk with milk newline).

As for your menu function, I can only assume bash doesn't like recursion much. One way around it would be:

Code:

function menu
{
while :
do
&lt;your menu code&gt;
done
}

This will create a never ending loop. Of course you could add a quit option and put it in as a condition to the while loop.

Edit: corrected half baked sentence count (a bit)