I want to construct variable name N_foo and N_bar and use their values in the following:
#!/bin/bash
N_foo=2
N_bar=3
for i in { "foo" "bar" }
do
for j in { 1..$(`N_$i`) }
do
echo $j
done
done
I want to use the values of N_foo and N_bar in the two inner loops and print out 1, 2 and 1, 2, 3, respectively. What's the correct syntax?
#!/bin/bash
N_foo=2
N_bar=3
for i in "foo" "bar"
do
key="N_${i}"
eval count='$'$key
for j in `seq 1 $count`
do
echo $j
done
done
eval is the classic solution to this sort of issue.
${!symbol} syntax you are asking about. Your edit is really a new question. If you decide not to open a new question for it, you should edit your own question instead of the answer.
seq 1 $count is enough.You can use the indirect variable reference operator:
Example
var="foo"
nfoo=1
ref=n${var}
echo $ref
echo ${!ref}
Which gives this output:
nfoo
1
#!/bin/bash
N_foo=2
N_bar=3
for i in "foo" "bar"
do
i2="N_$i"
seq 1 ${!i2}
done
seq is nonportable, not specified by POSIX nor included as part of bash.I ended up using the following code. It uses the parameter substitution technique (c.f. http://tldp.org/LDP/abs/html/parameter-substitution.html).
#!/bin/bash
N_foo=2
N_bar=3
for i in "foo" "bar"
do
j_max=N_$i
for (( j=1; j<=${!j_max}; j++ ))
do
echo $j
done
done
The ! is history expansion parameter (c.f. http://www.gnu.org/software/bash/manual/bashref.html#Special-Parameters). !j_max will be replaced by the most recent value set to j_max which is N_foo/N_bar in 1st/2nd iteration. It then calls ${N_foo}/${N_bar} which has value 2/3 in the 1st/2nd iteration.
{}from outer for and remove space next to the{}in inner loop.echo { 1..3 }does not produce what you want.