[solved] split or manipulate file based on matching entries in column

Question

I have a file in the following format:
500 [phone] [phone] 0001-0001
400 [phone] [phone] 0001-0001
100 [phone] [phone] 0001-0001
300 [phone] [phone] 0001-0001
200 [phone] [phone] 0001-0001
I'd like to end up with the following output where the entries in the first column are added if the 3rd column matches (I don't really care about the 2nd column)
In other words, 500 + 400 + 100 because the third column are all [phone]
1000 [phone] 0001-0001
500 [phone] 0001-0001
I've been playing around with trying to split the file based on the 3rd column and then walking through it line by line but I thought there must be a graceful one-liner to do this.
I have any Arch tools (repo or AUR) available to me and don't really mind the method as long as it's sane.
The file is always sorted reverse-numerically on the first column and then numerically on the third but I can do whatever it needs to get it working if required
Last edited by oliver (2014-05-16 13:04:56)

Answer 1

Thank you both Trilby and Saint0fCloud... I found a similar example online using awk and modified it a little:
awk '{array[$3]+=$1} END { for (f in array) {print array[f], f, $4, $5}}' $DATAFILE
It's getting closer but not quite there (and if I modify my test data it's clearer why)
datafile
500 [phone] [phone] 0003-0003
400 [phone] [phone] 0003-0003
100 [phone] [phone] 0003-0003
300 [phone] [phone] 0004-0004
200 [phone] [phone] 0004-0004
$ ./testfile
500 [phone] 0004-0004
1000 [phone] 0004-0004
If I can't get it going in awk, I'll switch over to the zsh method
edit
Not sure this is the most graceful but it seems to work
awk '{array[$3," " $4," " $5]+=$1} END { for (f in array) {print array[f], f}}' $DATAFILE
1000 [phone] 0003-0003
500 [phone] 0004-0004
If anyone sees anything horribly wrong with the logic, feel free to point it out.
Last edited by oliver (2014-05-15 17:10:45)