linux shell read command-Getting User Input Via Ke

2020-02-19 23:32:11

ref：http://bash.cyberciti.biz/guide/Getting_User_Input_Via_Keyboard

You can accept input from the keyboard and assign an input value to a user defined shell variable using read command.

read Command Syntax

read -p "Prompt" variable1 variable2 variableN

Where,

-p "Prompt" : Display prompt to user without a newline.
variable1 : The first input (word) is assigned to the variable1.
variable2 : The second input (word) is assigned to the variable2.

Handling Input

Create a script called greet.sh as follows:

#!/bin/bash
read -p "Enter your name : " name
echo "Hi, $name. Let us be friends!"

Save and close the file. Run it as follows:

chmod +x greet.sh
./greet.sh

Sample Outputs:

Enter your name : Vivek Gite
Hi, Vivek Gite. Let us be friends!

Examples

Try the following examples.

Multiple Input (number.sh)

#!/bin/bash
# read three numbers and assigned them to 3 vars
read -p "Enter number one : " n1
read -p "Enter number two : " n2
read -p "Enter number three : " n3
 
# display back 3 numbers - punched by user. 
echo "Number1 - $n1"
echo "Number2 - $n2"
echo "Number3 - $n3"

Display Domain Owner Information

A shell script to display the Internet domain name owner information (domain.sh):

#!/bin/bash
read -p "Enter the Internet domain name (e.g. nixcraft.com) : " domain_name
whois $domain_name

Timeout Input

You can time out read command using the -t option. It causes read to time out and return failure if a complete line of input is not read within TIMEOUT seconds. For example, if no input provided within 10 second, program will be aborted (domain2.sh):

#!/bin/bash
read -t 10 -p "Enter the Internet domain name (e.g. nixcraft.com) : " domain_name
whois $domain_name

Handling Passwords

The -s option causes input coming from a terminal do not be displayed on the screen. This is useful for password handling (readpass.sh):

#!/bin/bash
read -s -p "Enter Password  : " my_password
echo 
echo "Your password - $my_password"

Handling multiple values

Consider the following example:

read -p "Enter directory to delete : " dirname
echo "$dirname"

Sample outputs:

Enter directory to delete : foo bar /tmp/data
foo bar /tmp/data

The user supplied three values instead of one. The string is now made of three different fields. All three words are assigned to dirname using $IFS internal field separator. The $IFS determines how shell recognizes fields.

$IFS

To display default value of $IFS, enter:

echo "$IFS"

You will see a whitespace which is nothing but a space, a tab, and a newline (default). You can print actual values of IFS using the following command (see Here strings):

cat -etv <<<"$IFS"

Sample outputs:

 ^I$
$

Where,

$ - end of line i.e. newline
^I$ - tab and newline

But how do I use $IFS and read command together?

Create a variable called nameservers and give it total 3 values as follows (note all values are separated by a whitespace):

nameservers="ns1.nixcraft.net ns2.nixcraft.net ns3.nixcraft.net"

Display the value of a variable nameservers with echo command or printf command:

echo "$nameservers"

printf "%s" $nameservers

Now, you can simply split $nameservers using the read command as follows (see Here strings):

read -r ns1 ns2 ns3 <<< "$nameservers"

Where,

The read command reads input from $nameservers variable.
The default value of $IFS is used to assign values to three separate variables. Your input is broken into tokens using $IFS and assigned to three variables.
In other words, the IFS variable worked as token delimiter or separator.
The first token (ns1.nixcraft.net) is saved as the value of the first variable ($ns1)
The second token (ns2.nixcraft.net) is saved as the value of the second variable ($ns2).
The third token (ns3.nixcraft.net) is saved as the value of the third variable ($ns3).
To display the value of each variable use echo command or printf command as follows:

echo "DNS Server #1 $ns1"
echo " #2 $ns2"
echo " #3 $ns2"

OR use the printf command

printf "DNS Server #1 %s\n #2 %s\n #3 %s\n" $ns1 $ns2 $ns3

Sample outputs:

DNS Server #1 ns1.nixcraft.net
 #2 ns2.nixcraft.net
 #3 ns3.nixcraft.net

How do I change the IFS separator value?

Consider the following /etc/passwd line:

gitevivek:x:1002:1002::/home/gitevivek:/bin/sh

Assign the above line to a variable called pwd:

pwd="gitevivek:x:1002:1002::/home/gitevivek:/bin/sh"

Save the Internal Field Separator to a variable called old:

old="$IFS"

Set the Internal Field Separator to a colon (i.e. change the Internal Field Separator):

IFS=:

Read $pwd and generate tokens using $IFS and store them into respective fields:

read -r login password uid gid info home shell <<< "$pwd"
printf "Your login name is %s, uid %d, gid %d, home dir set to %s with %s as login shell\n" $login $uid $gid $home $shell

Sample outputs:

Your login name is gitevivek, uid 1002, gid 1002, home dir set to /home/gitevivek with /bin/sh as login shell

Finally, restore the Internal Field Separator value using $old:

IFS="$old"

Where,

: - act as token separator on $pwd i.e. the contents of the IFS variable are used as token delimiters.

login - Field # 1 is generated using the first token and is saved as the value of the first variable ($login)
password - Field # 2 is generated using the second token and is saved as the value of the second variable ($password)
uid - Field # 3 and so on...
gid - Field # 4
info - Field # 5
home - Field # 6
shell - Field # 7