Split string into 2, from 2nd occurance of a character in bash
 

I have a script, which is having a string as command line argument.

1= "ABC_MY_STING_anything_I_dont_care"

I need to have 2 variables

var1 = "ABC_MY"
var2 = "STRING_anything_I_dont_care"

Seperation is based on the "2nd" occurance of underscore.

I found answers on how to do it in python but I'm using bash.

Using tr doesnt look to solve this.

I tried using IFS=_ but that splits my string in a way I dont want, I need only 2 parts.

I know how to make regex but not getting how to split it this way.

anything_something_rest_of_the_things
[a-z A-Z 0-9]_[a-z A-Z 0-9]_


Background:
Currently I use mutt -s $1 to send email

But I need to split my email into 2 parts, var1 I'll use as subject and var2 I'll use as message body.

maybe something like

Thanks schneidz, but it says

Did what I exactly wanted, but still can anyone tell what -f1-2 meant for bash

d tells that delimiter is _ and what about f

Now that I have my subject in the s1 and body in s2. I cant find out a way to make s2 as the body. I wish to use mutt but, it expects file for making its body. Can anyone help on this.

Sorry, never used Linux, so someone might need to help edit this.
As far as regex-matching the text-strings, here's a sed example:
GNU version 4.0.7

sed -r -n 's/([^_]*_[^_]*)_(.*)/AnyTextHere \1/p' $1
sed -r -n 's/([^_]*_[^_]*)_(.*)/AnyTextHere \2/p' $1

They only output below vars if $1 meets your 2nd-occurrence requirement.
\1 = AllTextBefore2ndOccurenceOf_
\2 = AllTextAfterwards

You can always combine the \1\2 replacements, editing AnyTextHere with your needed apps (separated by &&) and then pipe to 'sed e' to auto-run the replacement text-string as a commandline.

IF $1 is of format: Mail_Subject_Body Text
Try something like below, but remove ;e for testing
Again, someone may need to help edit this for a bash shell

sed -r 's#([^_]*_[^_]*)_(.+)#echo -e "\2" | mutt -s "\1" user@earth.com#;e' $1

It creates text-lines as:
echo -e "Body Text" | mutt -s "Mail_Subject" user@earth.com

If that's not what you need, mod the replacement & test again.
Once in the correct format, add ;e back-in to execute the result.


** Edit **
Removed sed's -n and s///p option to make it simpler.
Its fine so long as $1 will always be in the right format

If not, give e a linematch like /@/e (for lines with email or whatever)
Or just pipe the changed $1's only with something like:

sed -n -r 's#([^_]*_[^_]*)_(.+)#echo -e "\2" | mutt -s "\1" user@earth.com#p' $1 | sed e

See the manual for cut
http://unixhelp.ed.ac.uk/CGI/man-cgi?cut

Delimiter is _ (underscores) in your case.
-f is the field list (as found using the delimiter you specified).
-f1-2 means fields 1 to 2.
-f3- means from field no 3 onwards.

OK

echo $s2 | mutt $s1 should be OK. If not, then substitute mutt with mailx.

OK

that looks overkill and overcomplicated for me. we can use IFS:
I think you use some script to construct this string and I suggest you to use different separator between the two fields. That will simplify this thing....