swapping columns

[joy_stat]

Hi,

I used the following command to swap the fist 2 columns of a file which contains around 100 columns.

gawk '{temp=$1;$1=$2;$2=temp}{print}' study_hm2_chr1.ped>temp1.ped
mv temp1.ped study_hm2_chr1.ped

I am trying to do this for a list of files in a loop but having no success. One method did swap the columns but deleted all the other columns which is not what I wanted. Is there a one-liner shell or perl command which will do this. This is what I am trying to do:

for i in {1..22}; do
SWAP 1st and 2nd columns of study_hm2_chr${i}.ped
done


Thanks,

-Joey

[John Rodkey]

Have you thought about just doing this?

for i in {1..22}; do
  awk '{print $2,$1}' study_hm2_chr${i}.ped > ped.$$
  mv ped.$$ study_hm2_chr${i}.ped
done

Seems like you're working too hard at it.

John

[joy_stat]

it does swap the first two columns for me but deletes the other columns...

[atreyu]

If you have a known, static number of columns (let's say five), you could simply include them, e.g.:

awk '{print $2,$1,$3,$4,$5}' study_hm2_chr${i}.ped > ped.$$

Or a more flexible solution might be something like:

cat study_hm2_chr1.ped|awk '{printf $2" "$1" "};{ for (x=3;x<=NF;x++) { printf "%s ", $x } };{print ""}'

I'm more of a perl slinger, though, so my awk code is probably questionable.

[joy_stat]

Thanks. I would really like an oneliner perl code but I don't know that much. In the interim, this seemed to work, though not elegant:

for i in {7..22}; do
gawk '{temp=$1;$1=$2;$2=temp}{print}' study_hm2_chr${i}.ped>temp${i}.ped;
mv temp${i}.ped study_hm2_chr${i}.ped;
done

[atreyu]

Well, it was more elegant than my hideous awk attempt.

How about this perl one-liner?

cat study_hm2_chr${i}.ped | perl -p -e 's/^(\w)[ \t]+(\w)(.*)$/$2 $1$3/'