Hi
I want to write a shell script that modifies a string - string1, so
that the last occurence of another string - string2 and anything after
that remains in string1. I believe I can use sed but cant figure out
how.
e.g:
s1="string1 string2 string3 string2 and more strings"
s2="string2"
output will be:
s1="string2 and more strings"

-Kay

2006-11-2, 10:45(-08), Kay:
> Hi
> I want to write a shell script that modifies a string - string1, so
> that the last occurence of another string - string2 and anything after
> that remains in string1. I believe I can use sed but cant figure out
> how.
> e.g:
> s1="string1 string2 string3 string2 and more strings"
> s2="string2"
> output will be:
> s1="string2 and more strings"
[...]

case $s1 in
*"$s2"*) s1=$s2${s1#*"$2"};;
esac

--
Stéphane

Kay wrote:
> Hi
> I want to write a shell script that modifies a string - string1, so
> that the last occurence of another string - string2 and anything after
> that remains in string1. I believe I can use sed but cant figure out
> how.
> e.g:
> s1="string1 string2 string3 string2 and more strings"
> s2="string2"
> output will be:
> s1="string2 and more strings"
>
> -Kay
>

If the strings contain no "/"s,


s1=$(echo "$s1" | sed "s/.*$s2/$s2/")

--
Regards,

---Robert

Robert Katz wrote:
> Kay wrote:
>> Hi
>> I want to write a shell script that modifies a string - string1, so
>> that the last occurence of another string - string2 and anything after
>> that remains in string1. I believe I can use sed but cant figure out
>> how.
>> e.g:
>> s1="string1 string2 string3 string2 and more strings"
>> s2="string2"
>> output will be:
>> s1="string2 and more strings"
>>
>> -Kay
>>
>
> If the strings contain no "/"s,
>
>
> s1=$(echo "$s1" | sed "s/.*$s2/$s2/")
>

I guess some punctuation charaters, like (") would cause trouble too.

--
Regards,

---Robert

2006-11-02, 20:39(+00), Robert Katz:
> Kay wrote:
>> Hi
>> I want to write a shell script that modifies a string - string1, so
>> that the last occurence of another string - string2 and anything after
>> that remains in string1. I believe I can use sed but cant figure out
>> how.
>> e.g:
>> s1="string1 string2 string3 string2 and more strings"
>> s2="string2"
>> output will be:
>> s1="string2 and more strings"
>>
>> -Kay
>>
>
> If the strings contain no "/"s,

And no . \ * $ [

> s1=$(echo "$s1" | sed "s/.*$s2/$s2/")

Note that it selects the rightmost $s2.

--
Stéphane

Stephane CHAZELAS wrote:
> 2006-11-02, 20:39(+00), Robert Katz:
>> Kay wrote:
>>> Hi
>>> I want to write a shell script that modifies a string - string1, so
>>> that the last occurence of another string - string2 and anything after
>>> that remains in string1. I believe I can use sed but cant figure out
>>> how.
>>> e.g:
>>> s1="string1 string2 string3 string2 and more strings"
>>> s2="string2"
>>> output will be:
>>> s1="string2 and more strings"
>>>
>>> -Kay
>>>
>> If the strings contain no "/"s,
>
> And no . \ * $ [

(") would cause trouble too.

>
>> s1=$(echo "$s1" | sed "s/.*$s2/$s2/")
>
> Note that it selects the rightmost $s2.

That's what she wanted (the rightmost $s2).


--
Regards,

---Robert

Kay wrote:
> Hi
> I want to write a shell script that modifies a string - string1, so
> that the last occurence of another string - string2 and anything after
> that remains in string1. I believe I can use sed but cant figure out
> how.
> e.g:
> s1="string1 string2 string3 string2 and more strings"
> s2="string2"
> output will be:
> s1="string2 and more strings"
>
> -Kay
>

For old shells I recommend expr:

s1=`expr x"$s1" : x".*\($s2.*\)"`



--
Michael Tosch @ hp : com

2006-11-02, 21:21(+00), Robert Katz:
> Stephane CHAZELAS wrote:
>> 2006-11-02, 20:39(+00), Robert Katz:
>>> Kay wrote:
>>>> Hi
>>>> I want to write a shell script that modifies a string - string1, so
>>>> that the last occurence of another string - string2 and anything after
>>>> that remains in string1. I believe I can use sed but cant figure out
>>>> how.
>>>> e.g:
>>>> s1="string1 string2 string3 string2 and more strings"
>>>> s2="string2"
>>>> output will be:
>>>> s1="string2 and more strings"
>>>>
>>>> -Kay
>>>>
>>> If the strings contain no "/"s,
>>
>> And no . \ * $ [
>
> (") would cause trouble too.

Why??

>>> s1=$(echo "$s1" | sed "s/.*$s2/$s2/")
>>
>> Note that it selects the rightmost $s2.
>
> That's what she wanted (the rightmost $s2).

Then

expr "x$s1" : "x.*\($s2.*\)"

should do (with same restrictions on the characters in $s2, and
$s1 must contain $s2).

--
Stéphane