I want to write a short script to ssh to different servers depending on the argument I provide - if it is a number, I want the number to be a part of the server name, if it is a letter I want it connect to a certain server. So far I have this:
#!/bin/bash
if [ $1 -eq ^[1,3-5]$ ]; then
ssh -X servername1$1
elif [ $1=h ]; then
ssh -X servername2
fi
but it keeps complaining that integer expression is expected and always tries to connect to servername2, no matter whether I give it a number or a letter.
I tried quotation marks around the first and the second comparison, I tried double equality sign in the second comparison - all to no avail. How can I make this work?
[ $1 -eq ^[1,3-5]$ ] is not bash syntax.
You can write like that:
#!/bin/bash
if [[ $1 =~ ^[1,3-5]$ ]]; then
echo "$1 integer"
elif [ $1 = h ]; then
echo "$1 letter"
fi
Output:
[sahaquiel#sahaquiel-PC Stackoverflow]$ ./numorlet.sh 1
1 integer
[sahaquiel#sahaquiel-PC Stackoverflow]$ ./numorlet.sh 3
3 integer
[sahaquiel#sahaquiel-PC Stackoverflow]$ ./numorlet.sh 10
[sahaquiel#sahaquiel-PC Stackoverflow]$ ./numorlet.sh h
h letter
[sahaquiel#sahaquiel-PC Stackoverflow]$ ./numorlet.sh hhhh
[sahaquiel#sahaquiel-PC Stackoverflow]$
Also, tell me please, if the letter will be only 'h', or it can be any?
If any, use [[ instead of [ also in elif condition, with regex (I think ^[a-zA-Z]$ will work)
I suppose you have a server list where you are taking the argument $1 from.
If yes, you will have to do some changes in the code.
#!/bin/bash
if [[ $1 -eq ^[1,3-5]$ ]]; then
ssh -X servername1$1 < /dev/null
elif [[ $1==h ]]; then
ssh -X servername2 < /dev/null
fi
SSH takes input from stdin by adding input from /dev/null you are telling ssh to take input from that specific file.
Using [[...]] tells that if has an expression.
you can read more about the [[..]] in this link "Confused about operators"
Also $1=h is an assignment not a comparison; == is used for comparison.
Related
I'm a little befuddled with a script I've been writing - and would appreciate some help!
This is one of those cases where each command seems to work fine on their own, but not so when put together into a script.
Here's a gist of what I'm trying to do:
input=$1
single_func () {
command "$input"
}
multi_func () {
xargs < $input -n 1 single_func
}
if [[ "$input" == name1* || name2* ]];
then
single_func
elif [[ -f "$input" ]];
then
multi_func
else
echo "exiting"
exit
fi
The idea here is - if the script is invoked with ./script.sh input, if will run if the input starts with name1 or name2, using single_func. If the input provided doesn't start with name1 or name2, and is a file containing a list of items, elif will run (reason for -f) using multi_fuc, which is just single_func running with xarg on the provided file.
The 'single_func' component runs on the command line fine on its own (command "input"), and the 'multi_func' component runs fine with a test file (xargs < testfile.txt -n 1 ./single_func.sh). But when I put them together as above and try to run them together, only the first 'if' part works correctly. When provided with a file or some nonsense line not containing name1 or name2, the script simply exits without returning anything.
For the curious, I'm running entrez direct commands within the single_func block.
What am I doing wrong?
You need to write:
if [[ "$input" = name1* || "$input" = name2* ]]; then
Otherwise, the right-hand side of your || tests whether name2* is a non-empty string, which it always unconditionally is, making the statement always true.
If you don't want to repeat yourself (and your real use case is complex enough you can't just change it to if [[ "$input" = name[12]* ]]), use a case statement instead:
case $input in
name1*|name2*) echo "Either name1 or name2 prefix found";;
*) echo "Neither prefix found";;
esac
I am trying to use a while loop to repeatedly prompt for a username during user registration until a valid username has been provided (i.e. it has NOT already been claimed).
Unfortunately, I've only been coding in bash for less than 70 hours now, so I wasn't familiar with how to do this. My initial thought was to use "goto" like in Windows batch scripts if there was a problem with the username, but apparently that's not a thing in bash and loops seemed to be the recommended route. I wasn't familiar with loops in bash so I did some research until I found a good answer on SO about how to do this. This isn't the exact post (I can't find it now), but I was looking at questions like this one.
The resulting code looks like this:
echo -e "Now, choose a unique username. No special characters, please."
uniqueuser=0 # Assume a valid username has not been chosen from the get-go
while [ "$uniqueuser" == "0" ]
do
read -p "Username: " username
lusername=`echo $username | sed y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/` #lowercase username
if [ "$lusername" == "admin" ] || [ "$lusername" == "administrator" ] || [ "$lusername" == "npstn" ] || [ "$lusername" == "phreak" ] || [ "$lusername" == "guest" ] || [ "$lusername" == "user" ] || [ "$lusername" == "sysop" ]
then
echo -e "That username is reserved. Please pick another username."
else
username=`php /home/com/encode.php "$username"`
available=`curl "https://example.com/redacted/redacted.php?key=$key&type=checkusername&username=$username" --silent`
if [ "$available" == "1" ]
then
uniqueuser=$((1)) # Username is unique and acceptable
else
echo -e "That username is taken. Please choose another one."
fi
fi
done <input.data # so that username variable persists outside of while loop
echo -e "That username is available... or was... now, it's yours!"
echo -e "On this board, we have no annoying password length or complexity requirements. That said, your password cannot be null, nor can it use the plus symbol. Choose wisely!"
When this part of the shell script is reached, I see:
Now, choose a unique username. No special characters, please.
/home/com/redacted.sh: line 4: input.data: No such file or directory
That username is available... or was... now, it's yours!
On this board, we have no annoying password length or complexity requirements. That said, your password cannot be null, nor can it use the plus symbol. Choose wisely!
The reason "input.data" is used is to get the $username variable out from the subshell and back into the main shell. However, that seems to be causing an error. How can I fix this?
You've misread the sources on which you're relying. Redirecting stdin from a file, as in:
count=0
while IFS= read -r line; do
(( ++count )) # or do something else
done < input.data
echo "$count" # will print number of lines in the file input.data
only prevents a subshell from being created when the alternative is redirecting stdin from a pipe. For example, you would indeed lose the content of variables you set in:
# DO NOT DO THIS; this is the practice the other answers were teaching how to avoid.
count=0
cat input.data | while IFS= read -r line; do
(( ++count )) # or otherwise modify data within the loop
done
echo "$count" # will print 0, because the loop was in a subshell
This change is not necessary, and not appropriate, when your input is from stdin and not from a file.
It's the cat input.data | that is a problem, for which < input.data is a replacement. If your original code had no cat input.data |, then you need not use < input.data to replace it.
Demonstrating that a while read loop can read from stdin, and retain its variables' values, without any such redirection:
count=0
prompt='Enter a string, or a blank line to stop: '
while IFS= read -r -p "$prompt" line && [[ $line ]]; do
(( ++count ))
printf 'Line %d: %s\n' "$count" "$line"
done
echo "Final counter is $count"
...keeps the same value for $count outside the loop it hand inside the loop, when a blank line is entered.
I have a problem with writing bash script. The problem is in comparison of strings. When I launch it, there's no errors. However in result, it is always changing the variable client.
So if for an example we have two lines in file
apple A
orange D
and if I give the who=A I expect to see in result apple, or if at D - orange
But no matter of what I choose A or D it is always giving me the result - orange
No matter of the strings, it always change the variable client, like ignoring the comparison. Please help.
while read line
do
IFS=" "
set -- $line
echo $2" "$who":"$1
if [[ "$2"="$who" ]]
then
echo "change"
client=$1
fi
done < $file
echo $client
So now I changed the code as in one of the comment below, but now the caparison always false therefore the variable client is always empty
while read -r line
do
#IFS=" "
#set -- $line
#echo $2" "$who":"$1
#if [[ "$2" = "$who" ]]
a="${line% *}"
l="${line#* }"
if [[ "$l" == "$who" ]]
then
echo "hi"
client="$a"
fi
done < $file
If you have data in a file with each line like apple D and you want to read the file and separate then items, the parameter expansion/substring extraction is the correct way to process the line. For example (note $who is taken from your problem statement):
while read -r line
do
fruit="${line% *}" # remove from end to space
letter="${line#* }" # remove from start to space
if [[ "$letter" == "$who" ]]
then
echo "change"
client="$fruit"
fi
done < $file
Short Example
Here is a quick example of splitting the words with parameter expansion/substring extraction:
#!/bin/bash
while read -r line
do
fruit="${line% *}"
letter="${line#* }"
echo "fruit: $fruit letter: $letter"
done
exit 0
input
$ cat dat/apple.txt
Apple A
Orange D
output
$ bash apple.sh <dat/apple.txt
fruit: Apple letter: A
fruit: Orange letter: D
Change if [[ "$2"="$who" ]] to
if [[ "$2" = "$who" ]]
spaces around =
Example (for clarification):
who=A
while read line
do
IFS=" "
set -- $line
echo $2" "$who":"$1
if [[ "$2" = "$who" ]]
then
echo "change"
client=$1
fi
done < file #this is the file I used for testing
echo $client
Output:
A A:apple
change
D A:orange
apple
For who=D:
A D:apple
D D:orange
change
orange
You do need spaces around that = operator.
However, I think you're facing yet another issue as you're trying to change the value of the client variable from inside the while loop (which executes in a subshell). I don't think that will work; see this quesion for details.
I have an issue in finding a part of string variable in another string variable, I tried many methods but none worked out..
for example:
echo -e " > Required_keyword: $required_keyword"
send_func GUI WhereAmI
echo -e " > FUNCVALUE: $FUNCVALUE"
flag=`echo $FUNCVALUE|awk '{print match($0,"$required_keyword")}'`;
if [ $flag -gt 0 ];then
echo "Success";
else
echo "fail";
fi
But it always gives fail though there are certain words in variable which matches like
0_Menu/BAA_Record ($required_keyword output string)
Trying to connect to 169.254.98.226 ... OK! Executing sendFunc GUI
WhereAmI Sent Function WhereAmI [OK PageName:
"_0_Menu__47__BAA_Record" ($FUNCVALUE output string)
As we can see here the BAA_Record is common in both of the output still, it always give FAIL
The output echo is
> Required_keyword: 0_Menu/BAA_Record
> FUNCVALUE:
Trying to connect to 169.254.98.226 ... OK!
Executing sendFunc GUI WhereAmI
Sent Function WhereAmI [OK]
PageName: "_0_Menu__47__BAA_Record"
Bash can do wildcard and regex matches inside double square brackets.
if [[ foobar == *oba* ]] # wildcard
if [[ foobar =~ fo*b.r ]] # regex
In your example:
if [[ $FUNCVALUE = *$required_keyword* ]]
if [[ $FUNCVALUE =~ .*$required_keyword.* ]]
Not sure if I understand what you want, but if you need to find out if there's part of string "a" present in variable "b" you can use simply just grep.
grep -q "a" <<< "$b"
[[ "$?" -eq 0 ]] && echo "Found" || echo "Not found"
EDIT: To clarify, grep searches for string a in variable b and returns exit status (see man grep, hence the -q switch). After that you can check for exit status and do whatever you want (either with my example or with regular if statement).
d is an internal server lookup tool I use.
I am looking to allow a user to input any number between 0 (or 1) and 9999 (let's call this userinput) and have it display the result of:
d $userinput (e.g. 1234)
Then manipulate the results of that lookup (below gets rid of everything but the IP address to ping later):
grep -E -o '(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)'`
I know I need to use the while true; do read $blah etc etc. I am just not familiar with read enough to format it properly and more importantly:
get it to prompt for a numerical input between 0-9999
The other answers have many flaws, because they check that the user didn't input a number outside of the range they want. But what if a user enters something that is not a number? their strategy is broken from the start.
Instead it's better to let go only when we're sure that the user entered a number which lies within the wanted range.
while :; do
read -ep 'Enter server number: ' number
[[ $number =~ ^[[:digit:]]+$ ]] || continue
(( ( (number=(10#$number)) <= 9999 ) && number >= 0 )) || continue
# Here I'm sure that number is a valid number in the range 0..9999
# So let's break the infinite loop!
break
done
The regex [[ $number =~ ^[[:digit:]]+$ ]] makes sure that the user only entered digits.
The clumsy (number=(10#$number)) part is here so that if the user enters a number that starts with a 0, bash would try to interpret it in radix 8 and we'd get a wrong result (e.g., if the user enters 010) and even an error in the case when a user enters, e.g., 09 (try it without this guard).
If you only want to prompt once and exit when the user inputs invalid terms, you have the logic:
read -ep 'Enter server number: ' number
[[ $number =~ ^[[:digit:]]+$ ]] || exit 1
(( ( (number=(10#$number)) <= 9999 ) && number >= 0 )) || exit 1
# Here I'm sure that number is a valid number in the range 0..9999
If you want to explain to the user why the script exited, you can use a die function as:
die() {
(($#)) && printf >&2 '%s\n' "$#"
exit 1
}
read -ep 'Enter server number: ' number
[[ $number =~ ^[[:digit:]]+$ ]] ||
die '*** Error: you should have entered a number'
(( ( (number=(10#$number)) <= 9999 ) && number >= 0 )) ||
die '*** Error, number not in range 0..9999'
# Here I'm sure that number is a valid number in the range 0..9999
<--edit-->
if all you want is the mechanic for prompting, try this:
echo -n "Enter server number:"
read userinput
then run validation checks on the input like this:
if [[ $userinput -lt 0 || $userinput -gt 9999 ]] # checks that the input is within the desired range
then
echo "Input outside acceptable range."
else
# insert your grep and ping stuff here
fi
<--end edit-->
on first read, i thought your problem sounded ideal for a wrapper script, so i was going to suggest this:
$ cat wrapper.sh
#!/usr/bin/bash
userinput=$1
if [[ $# != 1 ]] # checks that number of inputs is exactly one
then
echo "Too many inputs."
exit 2
elif [[ $userinput -lt 0 || $userinput -gt 9999 ]] # checks that the input is within the desired range
then
echo "Input outside acceptable range."
exit 3
fi
output=`d "$userinput"`
ping_address=`grep -E -o '(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)' <("$output")`
ping "$ping_address"
then call the script with like this:
$ wrapper.sh 1243
If you just want a number between two values, you can test their values:
read x
while [[ $x -lt 0 || $x -gt 9999 ]]; do
echo "bad value for x"
read x
done
echo "x=$x"
The read command doesn't prompt itself. Use a normal echo before to actually display a prompt. Use echo -n to not add a newline.