Assume that the Pid of active processes on my machines are 1000 and 2000.
I am trying to make an array in Linux such that
The command echo ${Pid_Current[0]} gives 1000 in output
The command echo ${Pid_Current[1]} gives 2000 in output
Here is my code:
declare -a Pid_Current
Pid_Current=$(ps -aF | tail -n +2 | awk '{print $2}')
However, instead of the desired output I explained above, I receive the following output:
echo ${Pid_Current[0]} gives 1000 2000 in output
echo ${Pid_Current[1]} gives nothing in output
Would you please advise me what part of my code is incorrect?
In bash array assignment is done by enclosing the expression in parenthesis, so to use array assignment you need to write:
Pid_Current=($(ps -aF | tail -n +2 | awk '{print $2}'))
Without parenthesis the result of the expression is assigned to Pid_Current[0]
Related
I have been trying to split up a string and putting it into an Array in Bash on my Mac without success.
Here is my sample code:
#!/bin/bash
declare -a allDisks
allDisksString="`ls /dev/disk* | grep -e 'disk[0-9]s.*' | awk '{ print $NF }'`"
#allDisksString="/dev/disk0s1 /dev/disk1s1"
echo allDisksString is $allDisksString
IFS=' ' read -ra allDisks <<< "$allDisksString"
echo allDIsks is "$allDisks"
echo The second item in allDisks is "${allDisks[1]}"
for disk in "${allDisks[#]}"
do
printf "Loop $disk\n"
done
And below is the output:
allDisksString is /dev/disk0s1 /dev/disk0s2 /dev/disk0s3 /dev/disk0s4 /dev/disk1s1
allDIsks is /dev/disk0s1
The second item in allDisks is
Loop /dev/disk0s1
Interesting if I execute the following in the Mac Terminal:
ls /dev/disk* | grep -e 'disk[0-9]s.*' | awk '{ print $NF }'
I get the following output
/dev/disk0s1
/dev/disk0s2
/dev/disk0s3
/dev/disk0s4
/dev/disk1s1
So I have also tried setting IFS to IFS=$'\n' without any success.
So no luck in getting a list of my drives into an array.
Any ideas on what I am doing wrong?
You're making this much more complicated than it needs to be. You don't need to use ls, you can just use a wildcard to match the device names you want, and put that in an array assignment.
#!/bin/bash
declare -a allDisks
allDisks=(/dev/disk[0-9]s*)
echo allDIsks is "$allDisks"
echo The second item in allDisks is "${allDisks[1]}"
for disk in "${allDisks[#]}"
do
printf "Loop $disk\n"
done
read only reads one line.
Use an assignment instead. When assigning to an array, you need to use parentheses after the = sign:
#!/bin/bash
disks=( $(ls /dev/disk* | grep -e 'disk[0-9]s.*' | awk '{ print $NF }') )
echo ${disks[1]}
I gather an array of numbers like this
values=`sed -n "3,6p" STAT_EE/table_EE'.tex' | awk '{ print $4}'`
val=`sed -n "8,12p" STAT_EE/table_EE'.tex' | awk '{ print $4}'`
and then I want to print this to variables in columns. I tried with echo -en "$values\t$val but it prints the two variables as a one single array I also tried with printf but it did not work.
Could you please help me with this concern? Thank's a lot
First, the variables that you have defined are not bash arrays; they are bash strings. To make arrays, enclose the expressions in parentheses:
values=(`sed -n "3,6p" STAT_EE/table_EE'.tex' | awk '{ print $4}'`)
val=(`sed -n "8,12p" STAT_EE/table_EE'.tex' | awk '{ print $4}'`)
To print them out in two columns:
for ((i=0;i<${#values[#]};i++))
do
echo "${values[i]} ${val[i]}"
done
or:
for i in $( seq 0 $((${#values[#]}-1)) )
do
echo "${values[i]} ${val[i]}"
done
Advanced Topics
Because the backtick notation is fragile and does not nest well, the %(...) construct is generally preferred. Also, in the sed-awk code above, sed is only being used to select and range of lines. That is something that awk does very well. So, the code can simplify to:
values=( $(awk 'NR>=3 && NR<=6{ print $4}' STAT_EE/table_EE.tex) )
val=( $(awk 'NR>=8 && NR<=12{ print $4}' STAT_EE/table_EE.tex) )
Hi I have written a c program that takes 3 integers as input:
./myprogram 1 2 3
and I am aiming to pipe data from a csv file into the input of the c program. I grab each line from the c program using:
for i in $(seq 1 `wc -l "test.csv" | awk '{print $1}'`); do sed -n $i'p' "test.csv"; done;
and then would like to pipe the output of this into my c program. I have tried doing:
for i in $(seq 1 `wc -l "test.csv" | awk '{print $1}'`); do sed -n $i'p' "test.csv"; done; | ./myprogram
however I get:
Line
bash: syntax error near unexpected token `|'
how do I pipe the output into my c program?
Thanks
It helps when you really try to understand error messages the shell gives you:
Line
bash: syntax error near unexpected token `|'
If you think about it, when you chain commands together in a pipeline, there is never a ; before a |, for example:
ls | wc -l
# and not: ls; | wc -l
Whatever comes after a ; is like an independent new command, as if you typed it on a completely new, clear command line. If you type | hello on a clear command line, you'll get the exact same error, because that's the exact same situation as ; | ... in your script, for example:
$ | hello
-bash: syntax error near unexpected token `|'
Others already answered this, but I also wanted to urge you to make other improvements in your script:
Always use $() instead of backticks, for example:
for i in $(seq 1 $(wc -l "test.csv" | awk '{print $1}')); ...
You didn't need the awk there, this would work just as well:
for i in $(seq 1 $(wc -l "test.csv")); ...
You could reduce your entire script to simply this, for the same effect:
./myprogram < test.csv
In the shell, it doesn't like an explicit line termination followed by a pipe (|). The pipe already delimits the commands. So you want:
for i in $(seq 1 `wc -l "test.csv" | awk '{print $1}'`); do sed -n $i'p' "test.csv"; done | ./myprogram
My bash script needs to read values from a properties file and assign them to a number of arrays. The number of arrays is controlled via configuration as well. My current code is as follows:
limit=$(sed '/^\#/d' $propertiesFile | grep 'limit' | tail -n 1 | cut -d "=" -f2- | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')
for (( i = 1 ; i <= $limit ; i++ ))
do
#properties that define values to be assigned to the arrays are labeled myprop## (e.g. myprop01, myprop02):
lookupProperty=myprop$(printf "%.2d" "$i")
#the following line reads the value of the lookupProperty, which is a set of space-delimited strings, and assigns it to the myArray# (myArray1, myArray2, etc):
myArray$i=($(sed '/^\#/d' $propertiesFile | grep $lookupProperty | tail -n 1 | cut -d "=" -f2- | sed 's/^[[:space:]]*//;s/[[:space:]]*$//'))
done
When I attempt to execute the above code, the following error message is displayed:
syntax error near unexpected token `$(sed '/^\#/d' $propertiesFile | grep $lookupProperty | tail -n 1 | cut -d "=" -f2- | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')'
I am quite sure the issue is in the way I am declaring the "myArray$i" arrays. However, any different approach I tried produced either the same errors or incomplete results.
Any ideas/suggestions?
You are right that bash does not recognize the construct myArray$i=(some array values) as an array variable assignment. One work-around is:
read -a myArray$i <<<"a b c"
The read -a varname command reads an array from stdin, which is provided by the "here" string <<<"a b c", and assigns it to varname where varname can be constructs like myArray$i. So, in your case, the command might look like:
read -a myArray$i <<<"$(sed '/^\#/d' $propertiesFile | grep$lookupProperty | tail -n 1 | cut -d "=" -f2- | seds/^[[:space:]]*//;s/[[:space:]]*$//')"
The above allows assignment. The next issue is how to read out variables like myArray$i. One solution is to name the variable indirectly like this:
var="myArray$i[2]" ; echo ${!var}
I need to know how many processes are running for a specific task (e.g. number of Apache tomcats) and if it's 1, then print the PID. Otherwise print out a message.
I need this in a BASH script, now when I perform something like:
result=`ps aux | grep tomcat | awk '{print $2}' | wc -l`
The number of items is assigned to result. Hurrah! But I don't have the PID(s). However when I attempt to perform this as an intermediary step (without the wc), I encounter problems. So if I do this:
result=`ps aux | grep tomcat | awk '{print $2}'`
Any attempts I make to modify the variable result just don't seem to work. I've tried set and tr (replace blanks with line-breaks), but I just cannot get the right result. Ideally I'd like the variable result to be an array with the PIDs as individual elements. Then I can see size, elements, easily.
Can anyone suggest what I am doing wrong?
Thanks,
Phil
Update:
I ended up using the following syntax:
pids=(`ps aux | grep "${searchStr}"| grep -v grep | awk '{print $2}'`)
number=${#pids[#]}
The key was putting the brackets around the back-ticked commands. Now the variable pids is an array and can be asked for length and elements.
Thanks to both choroba and Dimitre for their suggestions and help.
pids=($(
ps -eo pid,command |
sed -n '/[t]omcat/{s/^ *\([0-9]\+\).*/\1/;p}'
))
number=${#pids[#]}
pids=( ... ) creates an array.
$( ... ) returns its output as a string (similar to backquote).
Then, sed is called on the list of all the processes: for lines containing tomacat (the [t] prevents the sed itself from being included), only the pid is preserved and printed.
You may need to adjust the pgrep command (you may need or may not need the -f option).
_pids=(
$( pgrep -f tomcat )
)
(( ${#_pids[#]} == 1 )) &&
echo ${_pids[0]} ||
echo message
If you want to print the number of pids (with a message):
_pids=(
$( pgrep -f tomcat )
)
(( ${#_pids[#]} == 1 )) &&
echo ${_pids[0]} ||
echo "${#_pids[#]} running"
It should be noted that the pgrep utility and the syntax used are not standard.