Efficient way to loop through list

Shell Programming and Scripting
1

i have a variable that contains a list of files:

varA="/var/tmp/hello_0_myapp
/var/tmp/mello_1_myapp
/var/tmp/jello_2_myapp
/var/tmp/fello_3_myapp"

And i'm calling this variable in several functions:

col0 () {
        FILE=$(echo "${varA}" | egrep "_0_myapp" | sed "s~_0_myapp~~g")
                        eachfile=${FILE}
}

col1 () {
        FILE=$(echo "${varA}" | egrep "_1_myapp" | sed "s~_1_myapp~~g")
                        eachfile=${FILE}
}

col2 () {
        FILE=$(echo "${varA}" | egrep "_2_myapp" | sed "s~_2_myapp~~g")
                        eachfile=${FILE}
}

col3 () {
        FILE=$(echo "${varA}" | egrep "_2_myapp" | sed "s~_2_myapp~~g")
                        eachfile=${FILE}
}

As you can see here, inside each function, im going through each file in the "varA" variable which contains a very huge list of files. i'm echoing the varA variable several times.

how can this be done more efficiently?

btw, i only listed 4 functions here. but in the actual script, there are several functions. so i need this to be very efficient

2

Anything will be more efficient than the echo | grep | sed | awk | kitchen | sink you have now, but the basic shell for loop is intended for exactly this situation:

for X in $VAR
do
        echo "$X"
done

Do not quote $VAR, it depends on it being split on spaces or newlines.

It'd be helpful to see how you're actually using col0(), col1(), etc too.

3

Take a step back to where that variable is assigned. Where do the values come from? A file? The output of a "command substitution" (that could be redirected to a file as well)?
Why don't you read the values from that file in your functions?

4

I agree, the taken division into functions seems not optimal.
How often is each function called? A function makes most sense when it is called multiple times.
--
A technical simplification: the two commands

grep "_1_myapp" | sed "s~_1_myapp~~g"

can be done by only sed

sed -n  "s~_1_myapp~~gp"

The p modifier prints if a substitution took place.
The -n option suppresses the default print.
The g modifier only makes sense when you expect multiple _1_myapp per line.

5

OK, you seem to need the FILE variable to become a chopped-off filename depending on an integer 0 - 3. Given your variable contains

echo "$varA"
/var/tmp/hello_0_myapp
/var/tmp/mello_1_myapp
/var/tmp/jello_2_myapp
/var/tmp/fello_3_myapp

(whereever this might come from), you might get results very efficiently using an array (assuming bash ):

ArrA=( ${varA//_?_myapp/} )
echo ${ArrA[@]}
/var/tmp/hello /var/tmp/mello /var/tmp/jello /var/tmp/fello

and then assign FILE=${ArrA[2]} , for example, yielding

echo $FILE
/var/tmp/jello

. Would this come close to what you need? The number of "functions" would be (not quite, but almost) unlimited.

6

In bash or ksh93 you could use arrays, like so (keeping you col functions, but consolidating them into one:

varA="/var/tmp/hello_0_myapp
/var/tmp/mello_1_myapp
/var/tmp/jello_2_myapp
/var/tmp/fello_3_myapp"

oldIFS=$IFS
IFS=$'\n' 
arrA=( $varA )
IFS=$oldIFS

col() {
  eachfile=${arrA[$1]%%_*}
}

col 0; echo $eachfile
col 1; echo $eachfile
col 2; echo $eachfile
col 3; echo $eachfile

Output:

/var/tmp/hello
/var/tmp/mello
/var/tmp/jello
/var/tmp/fello