 
 
 
 
 
Linux Community Forums Software Command Line Select with long list
26 Apr 11
I'm writing a command line script, where I want the user to select a file name. The select command almost does the job, but the list of files is rather long, about 150.
Anyone know of a way of throwing something like less into the mix, to enable the user to scroll around the list before selecting?
Cheers.
27 Apr 11
dwclarke wrote:
I'm writing a command line script, where I want the user to select a file name. The select command almost does the job, but the list of files is rather long, about 150.
Anyone know of a way of throwing something like less into the mix, to enable the user to scroll around the list before selecting?
Cheers.
Select is rather strict and unflexible. You can relay on the usual buffer of any terminal (scrolling with SHIFT+PgUp/PgDown ).
You are better off writting your own menu though
Regards
27 Apr 11
dwclarke wrote:
I'm writing a command line script, where I want the user to select a file name. The select command almost does the job, but the list of files is rather long, about 150.
Anyone know of a way of throwing something like less into the mix, to enable the user to scroll around the list before selecting?
Cheers.
#!/bin/bash
listfiles(){
vars=`seq 1 150`
for var in $vars; do
echo $var
done
}
listfiles|more
until [ -n "$answer" ]; do
echo -n "Enter your selection: "
read answer
done
echo you selected $answer
just replace the vars= line with your 150 variables. btw, i used more instead of less - personal pref.
-bill
27 Apr 11
Thanks for that. I've moded it a bit, and have ended up with this:
#!/bin/bash
# Script for a user to select a file
echo "Remember the number of the file you want, then press 'q', and enter the number"
echo "Press any key to start"
# Wait for keypress
read -n 1 -s
# Show numerated file list
ls *.txt | cat -n | less
#Get users file selection number
until [ -n "$answer" ]; do
echo -n "Enter your selection: "
read answer
done
# Check entry is numeric
if echo $answer | grep "^[0-9]*$" > /dev/null
then
# check the number isn't bigger than the list of files
count=`ls *.txt | wc -l`
if [ $answer -le $count ]
then
ourfile=`ls *.txt | sed -n "$answer p"`
echo "You chose $ourfile"
else
echo "Number is too large"
fi
else
echo "$answer is not numeric"
fi
28 Apr 11
I think I learn more from this sub-forum than any of the others. Thanks to all that post code here. As someone who never had to write bash code, I like watching others discuss it.
29 Apr 11
If I may add some suggestions:
# Show numerated file list
ls *.txt | cat -n | less
Yo do not need to pipe anything here, just do
less -N *.txt
# Check entry is numeric
if echo $answer | grep "^[0-9]*$" > /dev/null
Again, you do not need to pipe anything. This checks if it contains non numeric characters: a2 or 34s will fail the test
if [[ ! $answer = *[!0-9]* ]]
# check the number isn't bigger than the list of files
count=`ls *.txt | wc -l`
Don't ever parse the output of ls!!!. Insted:
f=(*.txt)
count=${#f[@]}
WARNING: this does not check if there are no .txt files <--- it will fail in that case as it will return a one file (the * glob)
ourfile=`ls *.txt | sed -n "$answer p"`
echo "You chose $ourfile"
I'm not quite sure what's your intention here... ;)
I'm assuming you are using bash, aren't you?
You should avoid `command` as it is deprecated and not easy to nest (try nesting 3-4 things and you'll understand what I mean). Try $(command) instead ;)
If you are using bash.... you should use [ [ instead of [ . The [ command (yes, it is a command, not a built in shell option) is, by far, less powerfull than the built in [ [
Regards
The Linux Foundation is a non-profit consortium dedicated to the growth of Linux.
Join / Linux Training / Board
