Sed Delete Last Character

by on July 14, 2007 · 10 comments· LAST UPDATED July 14, 2009

in , ,

How do I delete a last character from input line under UNIX / Linux using sed tool? My sample scientific input data line is as follows and I'd like to get rid of last + character only:

379100.46+521870.11+363838.75+382869.59+

How do I used sed to delete last + character?

Use it as follows:
echo "379100.46+521870.11+363838.75+382869.59+" | sed '$s/.$//'
If you would like to calculate final result, enter:
echo "379100.46+521870.11+363838.75+382869.59+" | sed '$s/.$//' | bc -l

TwitterFacebookGoogle+PDF versionFound an error/typo on this page? Help us!

{ 10 comments… read them below or add one }

1 Mani October 18, 2010 at 11:47 am

Hi,

Please help me for this.

input : 379100.46+521870.11+363838.75+382869.59
output : 379100.46+521870.11
I would like remove last the character based last two delimiter. Please help me.

Reply

2 jupiter2005ster March 18, 2011 at 2:17 pm

Try it:

echo $text | sed -e ‘s/\(.*\)../\1/’

Reply

3 Konrad Höffner July 22, 2011 at 11:52 am

That was exactly what I was searching for, thanks!
Can you explain why it works, though? I use sed quite often but I don’t know what the

$s/.$

does, I always do

sed "s|text|replacement|g"

.

Reply

4 SilversleevesX November 14, 2011 at 4:46 am

Konrad,

The “$” (dollar sign; I usually call it “cash”) is a marker or flag that tells sed to jump to the end of a line and edit something in the vicinity of the last character in it. I believe its counterpart for the beginning of a line is “^” (up-pointing carat), but don’t quote me on that. These markers are part of the sed program itself, which can present a problem only when the characters they represent are the ones you want to “sed out” of a string — sometimes even escaping them doesn’t work in those instances.

Hope this helps.

BZT

Reply

5 SilversleevesX November 14, 2011 at 9:37 am

mandre on ivorde.ro offers other approaches here.
http://www.ivorde.ro/How_to_remove_first_last_character_from_a_string_using_SED-75.html

Evidently the ‘$s’ syntax at the beginning of the sed command isn’t common or even necessary. A simple

sed ‘s/.$//’

usually suffices.

(The above is from one of the comments, not ivorde’s primary text.)

BZT

Reply

6 Manas March 12, 2012 at 4:30 pm

I have file…called file1 in the below format

1,abc,2,te1,gem,
2,3,b,r,4
4,5,3,t,r,
9,8,y,6,5,
7,5,1,e,5

means at the end of some of the lines ,i don’t have comma(,)
now can i place comma in those lines so that my o/p looks like below

1,abc,2,te1,gem,
2,3,b,r,4,
4,5,3,t,r,
9,8,y,6,5,
7,5,1,e,5

Reply

7 Manas March 12, 2012 at 4:31 pm

sorry the o/p should be

1,abc,2,te1,gem,
2,3,b,r,4,
4,5,3,t,r,
9,8,y,6,5,
7,5,1,e,5,

Reply

8 Manas March 12, 2012 at 4:41 pm

how to add a new column(first column) with variable string value to an existing file

input file name ….filename1
format

col1,col2,col3,col4
1,2,3,4
5,4,3,2
m,g,f,r

o/p file format
col0,col1,col2,col3,col4
abc,1,2,3,4
abc,4,3,2,5
abc,8,7,6,3
abc,9,8,6,1

‘abc’ value will change based on the input to the INPUT file

can you pls help

Reply

9 randomhello August 15, 2012 at 6:37 am

379100.46+521870.11+363838.75+382869.59+

Deleting the + at the end is ok but what if there’s no + sign there, wouldn’t you delete the 9 then? and how to solve it so that only the + is deleted?

Reply

10 Konrad Höffner August 15, 2012 at 12:20 pm

@randomhello:

No problem, just replace the dot character ‘.’ which stands for “any character” by a ‘+’:

sed ‘s/+$//’

Reply

Leave a Comment

Tagged as: , , , , , , , , , ,

Previous Faq:

Next Faq: