Looking for the right tool for the job

[rogers42]

Hi Folks,

I have a shell script that processes some data and emails the results to a list of receipients. The email addresses are hard coded in the shell script.

I have written an awk utility to globally search and to replace certain email addresses with the new ones.

What I want to do is to update the original file (with the new email addresses) using my awk utility.

However, what I am getting is as follows

1, The original file is not updated
2, The utility prints the entire line where the change was made
3, If three changes are made on a line, the line is printed three times over.

Re-directing the output to a new file won't help because I will get the delta.

For starters, is awk the right tool for the job at hand ?

Can anybody suggest a work around (using awk) ? or a better tool to get the job done ?

PS: I don't know Perl. However, I am familiar with Korn shell and sed (a little bit)

Thanks in advance.

rogers42

 

[PHV]

What is your actual code ?

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[rogers42]

Hi,

The awk code is as follows

# Search for the old string and replace with a white space
gsub(/david@yahoo.com/, "");
gsub(/david@hotmail.com/, "");
gsub(/[Tt]amy@infinity.com/, "");
gsub(/[Aa]rthur@infinity.com/, "");

# Search for any of the following strings and replace with two new strings
gsub(/[Kk]athy@yahoo.com/, "howard@yahoo.com charles@yahoo.com");

The awk utility is invoked as follows

nawk -f search_and_replace.awk test1.txt

Thanks

rogers42

 

[AnotherAlan]

Why not just use sed;

sed 's/search string/new string/g' yourfile > newfile | mv newfile yourfile.

 

[rogers42]

Yup, sed is less painful. Thanks for the suggestion.

Can somebody please give me the (sed) syntax to replace an existing string with nothing (as opposed to a white space). Following is what I currently have

s/david@yahoo.com//g

The above code leaves a white hole.

Thanks in advance.

rogers42

 

[geirendre]

Think you must have a whitespace right before or after
the original mailadress, so eighter:

s/\ david@yahoo.com//g

or

s/david@yahoo.com\ //g

should remove it.

 

[PHV]

Perhaps this ?

s/ *david@yahoo.com *//g

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886

 

[rogers42]

Thanks geirendre and PHV. Both methods work well.

Correct me if I am wrong, but geirendre's code is trying to escape white spaces?

I am unclear as to how PHV's code fixes the problem. Are "*" used as wild characters to include white spaces ?

Thanks in advance

rogers42

 

[p5wizard]

In sed, * means: any number of the previous character or wildcard

so:
[tt]/ */[/tt] means: zero or more spaces
[tt]/ *john */[/tt] means: the word john surrounded at either end by zero or more spaces

for more info: see the man page for sed


HTH,

p5wizard

 

[rogers42]

Thanks for the explaination.

rogers42