Hi all,

I have a very large; delimited file. In vi I would like to replace:
CSACT_DY;AVG_UEACT1;uesPerActiveLinkSetSize_1;#;A
CSACT_DY;AVG_UEACT2;uesPerActiveLinkSetSize_2;#;A
CSACT_DY;AVG_UEACT3;uesPerActiveLinkSetSize_3;#;A

with:
CSACT_DY;AVG_UEACT1;Average uesPerActiveLinkSetSize_1;#;A
CSACT_DY;AVG_UEACT2;Average uesPerActiveLinkSetSize_2;#;A
CSACT_DY;AVG_UEACT3;Average uesPerActiveLinkSetSize_3;#;A

The only common thing is the AVG_. The last part is various lengths and ends in a mixture of numbers and capital letters

I tried
:%s/AVG_\(.*\);/AVG_\1;Average
but this matches to the last ;
Is there any way to change the default in vi to match to the first ;?

I have managed to hack around it by:
cat file.txt| awk -F\; '{print $1 ";" $2 "####" $3 ";" $4 ";" $5}' > file_2.txt
vi file_2.txt
:%s/AVG_\(.*\)####/AVG_\1;Average /
:%s/####/;

All weird and wonderful solutions welcome. Knowing vi, there must be about a hundred ways of doing this.

Cheers!

sed -e "s/;u/;Average u/" filename

or
%s/;u/;Average u
in vi

Sorry, I didn't make myself clear in my original posting. The snippet of the file was tiny and the first letter isn't always 'u'. Also, I only want to add in the word average when the previous word starts with AVG_.
So as an small example the text can contain things like:

CL_SUPD;URAUPDOC;uraUpdate_Other causes;#;A
CL_SUPD;URAUPPUU;uraUpdate_3_0_1 periodic URA update;#;A
CSACT_DY;AVG_UEACT1;uesPerActiveLinkSetSize_1;#;A
CSACT_DY;AVG_UEACT2;ActiveLinkSetSize;#;A
CSACT_DY;AVG_UEACT3;uesPerActiveLinkSetSize_2;#;A
CSACT_DY;BH_ACRALLCF;acRejections all causes;#;A
CSACT_DY;BH_ACRCDALF;acRejections_1_9_8 Code allocation failure;#;A

I would like to change only the AVG_ lines to:

CL_SUPD;URAUPDOC;uraUpdate_Other causes;#;A
CL_SUPD;URAUPPUU;uraUpdate_3_0_1 periodic URA update;#;A
CSACT_DY;AVG_UEACT1;Average uesPerActiveLinkSetSize_1;#;A
CSACT_DY;AVG_UEACT2;Average ActiveLinkSetSize;#;A
CSACT_DY;AVG_UEACT3;Average uesPerActiveLinkSetSize_2;#;A
CSACT_DY;BH_ACRALLCF;acRejections all causes;#;A
CSACT_DY;BH_ACRCDALF;acRejections_1_9_8 Code allocation failure;#;A

Done it! The part after AVG_ is always * characters or less, so this works:

:%s/AVG_\([A-Z0-9]\{1,8\};\)/AVG_\1Average /

Close, but no cigar
That worked for everything except where the AVG_**** was repeated in the third column. Eg:
RSUSG_DY;AVG_AVGDHT;AVG_AVGDHT;#;A
became:
RSUSG_DY;AVG_AVGDHT;AVG_AVGDHT;Average #;A

This approach is much better than mine above as it does not depend on the number of characters before the ;, so I had a fiddle with it and the following works:
sed -e "s/AVG_\([^;]*;\)\(.*\)/AVG_\1Average \2/"
Also taking out a bit of complexity, the following works:
sed -e "s/AVG_\([^;]*;\))/AVG_\1Average /"