AWK

AlexYZ · 02-23-2010, 06:03 PM

Hi

This is probably something really simple but my brains gone to mush :-(

I have a folder with only 24 files named <number>.dat (i.e. 4.dat, 6.dat and so on) where <number> is between 0 and 256. Each file has just two columns of data and nothing else.

I'm trying to combine all the second columns ($2) together. I've been fiddling around with getline and so far have

awk '{ getline ln < "6.dat" ; print ln" "$2 }' 4.dat

which takes file 4.dat and adds $2 from 6.dat, but I want a single command to take each $2 from every file and add them to (for example) 4.dat (having $1 from 4.dat is no problem). A command that takes every file in the folder and grabs $2 and places them in a common file would be ideal. Frankly I can work around if you combine both columns from every file.

Thanks,

Al

AnanthaP · 02-23-2010, 06:19 PM

Student?

awk can input all *.dat files in 1 shot - see the syntax. You can print the second variable and redirect the output to a file.

What have you got so far?

End

grail · 02-24-2010, 12:10 AM

how about:

cd into directory

for file in *; do
awk '{ print $2 }' $file >> combine.dat;
done

AlexYZ · 02-24-2010, 06:47 AM

Thanks for responses, eventually I found a workaround using paste, but it would be good to see what I'm missing. That said I'm out of time to look at this myself

AnanthaP - I've tried various combinations of the appropriate syntax (starting with wild card and moving on to [0-9]). For example

Code:

awk '{ getline ln < /^[1-9].dat/ ; print ln" "$2 }' 4.dat > combined.dat

which only takes $2 from 4.dat, while

Code:

awk '{ getline ln < /^[1-9].dat/ ; print ln" "$2 }' *.dat > combined.dat

combined $2 but all into a single column.

grail, thanks, although for me

Code:

for file in *; do
awk '{ print $2 }' $f >> combine.dat;
done

just hangs.

Alex

pixellany · 02-24-2010, 07:38 AM

Quote:

That said I'm out of time to look at this myself

Does that mean we are supposed to finish it?

In the suggestion from grail, I think the "$f" is supposed to be "$file"

AlexYZ · 02-24-2010, 08:09 AM

Quote:

Does that mean we are supposed to finish it?

nono not at all, quite the opposite, I don't want people to think I'd just given up trying and was waiting for an answer - I don't like leaving things but equally I only have a few days left with access to the system I'm working on and considering I've found a work around I'm just going to use that.

Replacing $f with $file makes it run, but it oupts everything to a single column rather than multiple ones (which is why I was using getline).

Many thanks