Variable b has a string. Awk retrieves a substring which I want to assign to variable c. This is what I did:
#!/bin/bash
b=$(llsubmit multiple.cmd)
echo $b | c=$(awk '{
ret=match($0,".in.")
rwt=match($0,"\" has")
rqt=rwt-(ret+4)
subs=substr($0,(ret+4),rqt)
}')
... but I get a blank output for echo $c:
You can't pipe into an assignment.
c=$(echo "$b" | awk '{
ret=match($0,".in.")
rwt=match($0,"\" has")
rqt=rwt-(ret+4)
subs=substr($0,(ret+4),rqt)
}')
(Notice also the quoting around $b.)
But your Awk script looks rather complex. And it doesn't produce any output. Should it print something at the end? Without access to sample output from llsubmit this is mildly speculative, but I'm guessing something like this could work:
c=$(echo "b" | sed -n 's/.*\(\.in\.[^"]*\)" has .*/\1/p')
(Notice also the backslashes to make the dots match literally.)
You should properly then use double quotes in echo "$c" too (unless you are completely sure that the output cannot contain any shell metacharacters).
... And, of course, very often you don't want or need to store results in a variable in shell scripts if you can refactor your code into a pipeline. Perhaps you are really looking for something like
llsubmit multiple.cmd |
sed -n 's/.*\(\.in\.[^"]\)" has .*/p' |
while read -r job; do
: things with "$job"
done
It's hard to tell from your question since you didn't provide sample input and expected output but is this what you're trying to do:
$ b='foo .in.bar has'
$ c="${b% has*}"
$ c="${c#*.in.}"
$ echo "$c"
bar
Related
I have this script that's designed to assign variables to commands that collect information about a system and then echo them back. This works very well for the first few commands, but the last one continues to return the value without "PRETTY_NAME=" stripped out of the output.
Is there some problem with this that I'm not seeing?
I have tried using grep to separate awk:
grep PRETTY_NAME /etc/*-release | awk -F '=' '{print $2}'
Using escaped quotes:
awk -F \"=\" '/PRETTY_NAME/ {print $2}' /etc/*-release
Whole block (edited somewhat for relevance)
declare -A CMDS=(
[primaryMacAddress]="cat /sys/class/net/$(ip route show default | awk '/default/ {print $5}')/address"
[primaryIpAddress]="hostname --ip-address"
[hostname]="hostname"
[osType]="awk -F '=' '/PRETTY_NAME/ {print $2}' /etc/*-release"
)
#This bit is actually nested in another function
for kpair in "${!CMDS[#]}" do
echo "$kpair=\"$( eval ${CMDS[$kpair]} )\""
done
Results when run from .sh file:
osType="PRETTY_NAME="Red Hat Enterprise Linux Server 7.4 (Maipo)""
expected:
osType=""Red Hat Enterprise Linux Server 7.4 (Maipo)""
When this command is run by itself, it seems to work as intended:
$ awk -F '=' '/PRETTY_NAME/ {print $2}' /etc/*-release
"Red Hat Enterprise Linux Server 7.4 (Maipo)"
Because your Awk command is specified in double quotes, interior dollar signs are subject to special treatment: the $2 is treated as a parameter substitution by your shell, and so the array element doesn't store the text $2 but rather its expansion. The Awk interpreter never sees the $2 syntax.
However, you have a second problem in your command dispatcher. Your eval command does not prevent word splitting:
eval ${CMDS[$kpair]}
you want this:
eval "${CMDS[$kpair]}"
without the quotes, your command is arbitrarily chopped into fields on whitespace. Then eval catenates the pieces together, using one space between them, and evaluates the resulting syntax. The difference can be demonstrated with the following example:
$ cmd="awk '/foo/ { print \$1\" \"\$2 }'"
$ echo 'foo a' | eval $cmd
foo a
$ echo 'foo a' | eval "$cmd"
foo a
We can just use echo to understand the issue:
$ echo $cmd
awk '/foo/ { print $1" "$2 }'
$ echo "$cmd"
awk '/foo/ { print $1" "$2 }'
The substitution of $cmd and the subsequent word splitting is done irrespective of any shell syntax that `cmd contains. We can see the pieces like this:
$ for x in $cmd ; do echo "<$x>" ; done
<awk>
<'/foo/>
<{>
<print>
<$1">
<"$2>
<}'>
When we execute eval $cmd, the above pieces are generated and re-combined by eval and evaluated. Needless to say, you don't want your command syntax to be chopped up and re-combined like this; who knows what sort of hidden bug will arise. It may be okay for the commands you have now, but as a generic command dispatch mechanism, it is flawed.
I have been working on a KornShell (ksh) script where I am struck with an error with sed expression.
I have a file named abc with 100 entries and I want to assign every 8th argument in every line of abc file to a variable.
I have used something like this.
#!/bin/ksh
typeset -i x=1
while read line ; do
var1=$(sed -n '$xp' abc.txt | awk '{print $8}')
print $var1
x="$x+1"
done < abc.txt
exit
I want to refer to variable x as the line number, but I am getting an error with sed expression in referencing x variable. Please help me out.
Your quotes are wrong. Anything in single quotes is a verbatim string; if you want variable interpolation, you need to use double quotes (or, in very special circumstances, no quoting at all).
You might as well refactor everyting into Awk, too. Trivially,
var1=$(awk -v n="$x" 'NR==n{ print $8 }' abc.txt)
However, the main loop reading the whole file again just to get one line out of it is highly inefficient. Maybe you want something like
awk '{ print NR, $8 }' abc.txt |
while read x var1; do
print "$var1"
# presumably do something with $x too?
done
I wanted to print the name from the entire address by shell scripting. So user1#12.12.23.234 should give output "user1" and similarly 11234#12.123.12.23 should give output 11234
Reading from the terminal:
$ IFS=# read user host && echo "$user"
<user1#12.12.23.234>
user1
Reading from a variable:
$ address='user1#12.12.23.234'
$ cut -d# -f1 <<< "$address"
user1
$ sed 's/#.*//' <<< "$address"
user1
$ awk -F# '{print $1}' <<< "$address"
user1
Using bash in place editing:
EMAIL='user#server.com'
echo "${EMAIL%#*}
This is a Bash built-in, so it might not be very portable (it won't run with sh if it's not linked to /bin/bash for example), but it is probably faster since it doesn't fork a process to handle the editing.
Using sed:
echo "$EMAIL" | sed -e 's/#.*//'
This tells sed to replace the # character and as many characters that it can find after it up to the end of line with nothing, ie. removing everything after the #.
This option is probably better if you have multiple emails stored in a file, then you can do something like
sed -e 's/#.*//' emails.txt > users.txt
Hope this helps =)
I tend to use expr for this kind of thing:
address='user1#12.12.23.234'
expr "$address" : '\([^#]*\)'
This is a use of expr for its pattern matching and extraction abilities. Translated, the above says: Please print out the longest prefix of $address that doesn't contain an #.
The expr tool is covered by Posix, so this should be pretty portable.
As a note, some historical versions of expr will interpret an argument with a leading - as an option. If you care about guarding against that, you can add an extra letter to the beginning of the string, and just avoid matching it, like so:
expr "x$address" : 'x\([^#]*\)'
I have two strings and I want to split with space and use them two by two:
namespaces="Calc Fs"
files="calc.hpp fs.hpp"
for example, I want to use like this: command -q namespace[i] -l files[j]
I'm a noob in Bourne Shell.
Put them into an array like so:
#!/bin/bash
namespaces="Calc Fs"
files="calc.hpp fs.hpp"
i=1
j=0
name_arr=( $namespaces )
file_arr=( $files )
command -q "${name_arr[i]}" -l "${file_arr[j]}"
echo "hello world" | awk '{split($0, array, " ")} END{print array[2]}'
is how you would split a simple string.
if what you want to do is loop through combinations of the two split strings, then you want something like this:
for namespace in $namespaces
do
for file in $files
do
command -q $namespace -l $file
done
done
EDIT:
or to expand on the awk solution that was posted, you could also just do:
echo $foo | awk '{print $'$i'}'
EDIT 2:
Disclaimer: I don not profess to be any kind of expert in awk at all, so there may be small errors in this explanation.
Basically what the snippet above does is pipe the contents of $foo into the standard input of awk. Awk reads from it's standard in line by line, separating each line into fields based on a field separator, which is any number of spaces by default. Awk executes the program that it is given as an argument. In this case, the shell expands '{ print $'$1' }' into { print $1 } which simply tells awk to print field number 1 of each line of its input.
If you want to learn more I think that this blog post does a pretty good job of describing the basics (as well as the basics of sed and grep) if you skip past the more theoretical stuff at the start (unless you're into that kind of thing).
I wanted to find a way to do it without arrays, here it is:
paste -d " " <(tr " " "\n" <<< $namespaces) <(tr " " "\n" <<< $files) |
while read namespace file; do
command -q $namespace -l $file
done
Two special usage here: process substitution (<(...)) and here strings (<<<). Here strings are a shortcut for echo $namespaces | tr " " "\n". Process substitution is a shortcut for fifo creation, it allows paste to be run using the output of commands instead of files.
If you are using zsh this could be very easy:
files="calc.hpp fs.hpp"
# all elements
print -l ${(s/ /)files}
# just the first one
echo ${${(s/ /)files}[1]} # just the first one
I am a newbie in Bash and I am doing some string manipulation.
I have the following file among other files in my directory:
jdk-6u20-solaris-i586.sh
I am doing the following to get jdk-6u20 in my script:
myvar=`ls -la | awk '{print $9}' | egrep "i586" | cut -c1-8`
echo $myvar
but now I want to convert jdk-6u20 to jdk1.6.0_20. I can't seem to figure out how to do it.
It must be as generic as possible. For example if I had jdk-6u25, I should be able to convert it at the same way to jdk1.6.0_25 so on and so forth
Any suggestions?
Depending on exactly how generic you want it, and how standard your inputs will be, you can probably use AWK to do everything. By using FS="regexp" to specify field separators, you can break down the original string by whatever tokens make the most sense, and put them back together in whatever order using printf.
For example, assuming both dashes and the letter 'u' are only used to separate fields:
myvar="jdk-6u20-solaris-i586.sh"
echo $myvar | awk 'BEGIN {FS="[-u]"}; {printf "%s1.%s.0_%s",$1,$2,$3}'
Flavour according to taste.
Using only Bash:
for file in jdk*i586*
do
file="${file%*-solaris*}"
file="${file/-/1.}"
file="${file/u/.0_}"
do_something_with "$file"
done
i think that sed is the command for you
You can try this snippet:
for fname in *; do
newname=`echo "$fname" | sed 's,^jdk-\([0-9]\)u\([0-9][0-9]*\)-.*$,jdk1.\1.0_\2,'`
if [ "$fname" != "$newname" ]; then
echo "old $fname, new $newname"
fi
done
awk 'if(match($9,"i586")){gsub("jdk-6u20","jdk1.6.0_20");print $9;}'
The if(match()) supersedes the egrep bit if you want to use it. You could use substr($9,1,8) instead of cut as well.
garph0 has a good idea with sed; you could do
myvar=`ls jdk*i586.sh | sed 's/jdk-\([0-9]\)u\([0-9]\+\).\+$/jdk1.\1.0_\2/'`
You're needing the awk in there is an artifact of the -l switch on ls. For pattern substitution on lines of text, sed is the long-time champion:
ls | sed -n '/^jdk/s/jdk-\([0-9][0-9]*\)u\([0-9][0-9]*\)$/jdk1.\1.0_\2/p'
This was written in "old-school" sed which should have greater portability across platforms. The expression says:
don't print lines unless they match -n
on lines beginning with 'jdk' do:
on a line that contains only "jdk-IntegerAuIntegerB"
change it to "jdk.1.IntegerA.0_IntegerB"
and print it
Your sample becomes even simpler as:
myvar=`echo *solaris-i586.sh | sed 's/-solaris-i586\.sh//'`