UNIX for Dummies Questions & Answers

I want to do a search and replace in a file but there may or may not be one space, two or none.

i.e.:

test 3,1
test   3,1
test,3,1
test   3,1

The output I will be looking for is this:

test 3,2
test  3,2
test,3,2
test   3,2

I have to match the first word, as the same numbers may appear through the file and I have to maintain the same number of spaces.

When I was using grep, I could use the dollar $ as a wildcard.

Can sed do this in some way? Or is there another way around it?

Thanks

I think that
sed 's/,1/,2/'
is all you need.

The above approach will work just fine. You could also use vi and simply do a global search and replace. vi the file and hit your colon key and then enter the following (thus your whole command will look like the following) ":%s/1/2/g"

I wont be using vi as I will be automating this from the script. As stated, I cannot subsitute on the comma one (,1) as it will appear in other places in the file.

I figure that I will just try this:

sed s/"test 3,1"/"test 3,2"/g
sed s/"test  3,1"/"test  3,2"/g
sed s/"test   3,1"/"test   3,2"/g
sed s/"test,3,1"/"test,3,2"/g

This is more processes than I want for the script, but for the moment it will have to do...

You can always do several s commands with a single sed process. But I think you are looking for something like this...

sed 's/\(test[ ,]*3,\)1/\12/'

But if it isn't asking too much, can you please provide a breakdown of that solution.

Thanks again.

Code:

1 sed 's/\(test[ ,]*3,\)1/\12/'
2 sed 's/                /   /'  
3 sed 's/\(           \) /\1 /'
4 sed 's/\(test[ ,]*3,\) /\1 /'
5 sed 's/\(test[ ,]*3,\)1/\12/'

Line 1 and line 5 are the finished product.

Line 2 shows the basic structure of the sed substitute command. Whatever is matched in the first field is replaced by the second field.

In line 3 I have added the "\(" and "\)". Everything between them is a remembered sub-field. And I can use the remembered sub-field in the result by using "\1". I could also have other remembered sub-fields and then I would use \2 \3 etc.

In line 4 I have added what I am remembering. The only tricky part might be "[ ,]*". The [ ,] matches either a space or a comma. The * says I may have zero or more.

In line 5 I added the last two characters. That 1 is there so I can throw it away by not remembering it. The 2 is what I needed to add.