UNIX for Dummies Questions & Answers

Sorry, I accidentally posted this in the advance forum first...

Hi, I'd like to use cut to remove the last field from a record

the original record is comma delimited with null as the 19th field. Example: field18,,field20

if I use
cut -d, -f1-19 file
the record ends with
field18,

but I'd like field19 to be shown also, like
field18,,

is there something I am missing, or should I just append a comma to the end of the record after removing field20 (the last field)

Thanks in advance,
kristy

you should look into a regular exprestion. to take what you want.

The cut utility would work great. if you wanted the extra null field but you dont. so i would suggest the following.

Code:

cut -d, -f1-19 file |sed 's/,$/,,/'

the "$" means end of line. so it will search for ",," @ the end of each line givein to it and remove the 2 commas and replace it w/ 1 comma.

Thanks, Optimus. This replaces the single comma with a double at the end of each record, which is exactly what I was looking for. However, a coworker tells me that we should be using awk:

awk -F, '{print $1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17,$18,$19}' file > newfile

now, this doesn't make sense to me because it is unwieldy and it prints white space between fields, instead of retaining its comma delimited structure. Even if we manipulated this to suit our needs, is there a clear advantage of using awk over cut piped to sed?

the only differance you are looking at is speed, and w/ your awk statement you are specifing exactly which fields you want to keep and not keep. you can do w/ the same thing w/ cut saying which fields you dont want.

Personaly i would take my method vs the current awk statement due to not to be mean but its an eye soar. When i first started doing scripting i did it the long way (that awk statement) now i find i am able ot do the exact same thing but shorter and quicker, and much faster to read thru for errors. IMHO

Thanks again. I was thinking the same thing.