elif conditional statement not working - linux

Question

I have this file as:
The number is %d0The number is %d1The number is %d2The number is %d3The number is %d4The number is %d5The number is %d6The...
The number is %d67The number is %d68The number is %d69The number is %d70The number is %d71The number is %d72The....
The number is %d117The number is %d118The number is %d119The number is %d120The number is %d121The number is %d122
I want to pad it like:
The number is %d0 The number is %d1 The number is %d2 The number is %d3 The number is %d4 The number is %d5 The number is %d6
The number is %d63 The number is %d64 The number is %d65 The number is %d66 The number is %d67 The number is %d68 The number is %d69
d118The number is %d119The number is %d120The number is %d121The number is %d122The number is %d123The number is %d124The
Please tell me how to do it through shell script
I am working on Linux

Answer 1

Edit:
This single command pipeline should do what you want:
sed 's/\(d[0-9]\+\)/\1 /g;s/\(d[0-9 ]\{3\}\) */\1/g' test2.txt >test3.txt
# ^ three spaces here
Explanation:
For each sequence of digits following a "d", add three spaces after it. (I'll use "X" to represent spaces.)
d1 becomes d1XXX
d10 becomes d10XXX
d100 becomes d100XXX
Now (the part after the semicolon), capture every "d" and the next three character which must be digits or spaces and output them but not any spaces beyond.
d1XXX becomes d1XX
d10XXX becomes d10X
d100XXX becomes d100
If you want to wrap the lines as you seem to show in your sample data, then do this instead:
sed 's/\(d[0-9]\+\)/\1 /g;s/\(d[0-9 ]\{3\}\) */\1/g' test2.txt | fold -w 133 >test3.txt
You may need to adjust the argument of the fold command to make it come out right.
There's no need for if, grep, loops, etc.
Original answer:
First of all, you really need to say which shell you're using, but since you have elif and fi, I'm assuming it's Bourne-derived.
Based on that assumption, your script makes no sense.
The parentheses for the if and elif are unnecessary. In this context, they create a subshell which serves no purpose.
The sed commands in the if and elif say "if the pattern is found, copy hold space (it's empty, by the way) to pattern space and output it and output all other lines.
The first sed command will always be true so the elif will never be executed. sed always returns true unless there's an error.
This may be what you intended:
if grep -Eqs 'd[0-9]([^0-9]|$)' test2.txt; then
sed 's/\(d[0-9]\)\([^0-9]\|$\)/\1 \2/g' test2.txt >test3.txt
elif grep -Eqs 'd[0-9][0-9]([^0-9]|$)' test2.txt; then
sed 's/\(d[0-9][0-9]\)\([^0-9]\|$\)/\1 \2/g' test2.txt >test3.txt
else
cat test2.txt >test3.txt
fi
But I wonder if all that could be replaced by something like this one-liner:
sed 's/\(d[0-9][0-9]?\)\([^0-9]\|$\)/\1 \2/g' test2.txt >test3.txt
Since I don't know what test2.txt looks like, part of this is only guessing.