for-loop in bash

vanKey · 06-23-2022, 06:03 AM

Hello all,

I have to write a for-loop, that will go through every line and replace, if present, the second minus sign by e-.

content of the test.dat-file:

Code:

0.500000
-0.50000
-1.000-1
-0.50000

content of the corrected file:

Code:

0.500000
-0.50000
-1.000e-1
-0.50000

I tried

Code:

awk 'BEGIN{FS=OFS=";"} {gsub(/\-/, "e-", $1)} 1' test.dat

,
but this will replace all minus signs. I only want to convert the second minus sign of each line.

Has anyone an idea, how to to that?

Kind regards
vanKey

NevemTeve · 06-23-2022, 06:31 AM

Code:

echo '012-345-678-9' |
sed 's/^\([^-]*\)-\([^-]*\)-\(.*\)$/\1-\2e-\3/'

result:

Code:

012-345e-678-9

pan64 · 06-23-2022, 06:31 AM

yes, just it is not bash (but awk).
Anyway, you need to use a different regexp (for example):

Code:

gsub(/0[-]/, "0e-", $1)

(not tested)

vanKey · 06-23-2022, 06:49 AM

Quote:

Originally Posted by pan64

yes, just it is not bash (but awk).
Anyway, you need to use a different regexp (for example):

Code:

gsub(/0[-]/, "0e-", $1)

(not tested)

The problem is, that it won't work for

Code:

0.500000
-0.50000
-1.001-1
-1.002-1
-0.50000

Kind regards
vanKey

pan64 · 06-23-2022, 07:10 AM

probably you made a mistake. It works for me. From the other hand you can modify that regexp to fit your needs.

vanKey · 06-23-2022, 07:20 AM

Quote:

Originally Posted by pan64

probably you made a mistake. It works for me. From the other hand you can modify that regexp to fit your needs.

yeah, i changed my post to the following:

It does not work for

Code:

0.500000
-0.50000
-1.001-1
-1.002-1
-0.50000

Kind regards
vanKey

vanKey · 06-23-2022, 07:47 AM

Code:

gsub(/1[-]|2[-]/, "1e-", $1)

could do, but it still only replaces everything with "1e-" then.

teckk · 06-23-2022, 02:53 PM

Basic Examples, not the whole answer. You finish it.

Code:

a=(
0.500000
-0.50000
-1.001-1
-1.002-1
-0.50000
)

for i in "${a[@]}"; do
    echo "Last 2 are : "${i: -2}""
done

for i in "${a[@]}"; do
    if [[ "${i: -2}" =~ ^- ]]; then
        echo "These end in -? : "$i""
    fi
done

for i in "${a[@]}"; do
    if [[ "${i: -2}" =~ ^- ]]; then
        echo "${i%??}"
    else
        echo "$i"
    fi
done

for i in "${a[@]}"; do
    if [[ "${i: -2}" =~ ^- ]]; then
        echo "${i%??}e-"
    else
        echo "$i"
    fi
done

astrogeek · 06-23-2022, 06:53 PM

Quote:

Originally Posted by vanKey

Code:

gsub(/1[-]|2[-]/, "1e-", $1)

could do, but it still only replaces everything with "1e-" then.

Rather than telling others that their suggestions do not work for some case, try to understand why they do not work and from that point ask yourself how they might be modified to work for the case at hand and try to make the change! You will learn much more and more quickly that way!

When attempting to write more complex regular expressions, I find it very useful to write out a description of what I want in concise, plain language, then using that as a sort of specification try to implement the behavior in a regular expression.

Again, please review the Site FAQ for guidance in asking well formed questions. Especially visit the link from that page, How to Ask Questions the Smart Way for discussion of things to consider when asking others for help.

grail · 06-23-2022, 11:13 PM

As others have said, your solution is in awk and not bash and there is no for loop, so the starting point would be to right your question with the correct information.

As you are using awk, I would bookmark http://www.gnu.org/software/gawk/man...ode/index.html

Now based on your first attempt and example:

Code:

awk 'BEGIN{FS=OFS=";"} {gsub(/\-/, "e-", $1)} 1' test.dat

Code:

0.500000
-0.50000
-1.000-1
-0.50000

My first question would be, why are you setting FS and OFS to something that does not appear in your input?

Based on your requirement:
[quote]go through every line and replace, if present, the second minus sign by e-.[/code]
I assume this means we only want to worry about lines with 2 minus signs, so the following would be ignored:

Code:

1.000-1

The above being the case, your main function call should have something telling it to firstly, only act on lines with 2 minus signs:

Code:

awk 'split($0,a,"-") == 3{<do change here>}' file

Then the other nice thing you could look at in the manual is the gensub command:

Code:

gensub(regexp, replacement, how [, target])

Focus on the "how":

Quote:

If how is a string beginning with ‘g’ or ‘G’ (short for “global”), then replace all matches of regexp with replacement. Otherwise, treat how as a number indicating which match of regexp to replace.

MadeInGermany · 06-24-2022, 12:12 AM

sed has the /N modifier (Nth match)

Code:

sed 's/-/e-/2'

MadeInGermany · 06-26-2022, 03:15 AM

A sed solution that replaces the first dash not at the beginning:

Code:

sed 's/\(.\)-/\1e-/'

The dot is a wildcard; because it is substituted it needs to be captured with \( \) and restored with \1

You can precise the wildcard to match only a digit, then an existing e would not match (and insert another e).

Code:

sed 's/\([0-9]\)-/\1e-/'