Unix Newbie Question

Beta Ray Bill

Hi All

I have scripting question.
I have a file "out.txt"

which is generated by another script

the file contains the following

my_identifier8859574

logout

The number is generated in the script and I have put the my_identifier bit in front of it as a unique identifier

I now have another script which uses this file, It needs the number Only that comes after my_identifier

I havn't a breeze how to do this.
I assume i need to use awk

something like

VALUE=cat out.txt | grep my_identifier | awk .....

any suggestions?
Cheers
Graham

backboiler

If the text is actually "my_identifier" (13 characters long) then you can use

head -1 out.txt | cut -c14-

That gives the 14th and subsequent characters of the first line of the out.txt file. If the text is variable length then add a space or other divider between the fields e.g. "my_identifier:733739" or "37efh8:73654343" use ":" as the field separator, then you can use

head -1 out.txt | cut -f2 -d':'

That takes the second field (-f instead of -c) and says the delimiter is a colon. You can use any character.

If you want to put the result in a shell variable then enclose the whole thing in "$(...)", i.e.

variable=$(head -1 out.txt | cut -f2 -d':')

Beta Ray Bill

Im not sure this will 100% work

As if the script that generates the out.txt throws up and oracle error its is very likey that a ':' will occur in the first line and hence the script which uses out.txt will be processing incorrect data.

that is why I user my_identifier as a unique identifier that could not occur if the script which generates out.txt throws up an error

for example if an ORA-0600:some_problem occured the second script would pull back the wrong data

Graham

backboiler

I'm not really familiar with awk's syntax but perl is supposedly similar. In perl it'd be

perl -ne 'print $1 if /my_identifier(\d+)/' out.txt

(assuming that the bit you're looking for is numeric, "\d+" can change to ".+" for all characters to the end of the line or "\S" (capital 'S') for non-whitespace characters).

Beta Ray Bill

figured this one out

its

VALUE="$(fgrep my_identifier out.txt)" VALUE="${VALUE#my_identifier}"

backboiler

Very tidy altogether!

Snowbat

grep 'my_identifier' out.txt | sed 's/my_identifier//'