UNIX for Dummies Questions & Answers

Dear all,

I 'd like to create a new txt file using the old file. For example, in old file, if count=2 then in new file, repeat that row twice, with the only difference is: on the first row, 'start' column contains the 1st apart of the 'start' in the old file; while in the 2nd row, the 'start' column contains the 2nd part of the 'start' coulmn in the old file.

Thanks a lot!

old file example:
chr gene_name count start
1_q OMA1 2 43210,243007,
1_q TFR1 3 429310,2412007,23366,
1_q FTOMA1 5 3290,243317,24333,2433136,24334,
...

new file I want:
chr gene_name count start
1_q OMA1 2 43210
1_q OMA1 2 243007
1_q TFR1 3 429310
1_q TFR1 3 2412007
1_q TFR1 3 23366
1_q FTOMA1 5 3290
1_q FTOMA1 5 243317
1_q FTOMA1 5 24333
1_q FTOMA1 5 2433136
1_q FTOMA1 5 24334
...

Using Perl :

Code:

cat old_file | perl -e '
while(<>){
 my ($chr,$gene,$count,$start) = split ;
 print "$chr\t$gene\t$count\t$_\n" foreach (split(/,/,$start));
 }
 '  > new_file

Using awk :

Code:

awk '{
split($4,arr,",");
for (tmp in arr) if (arr[tmp] !~ /^\s*$/) printf $1"\t"$2"\t"$3"\t"arr[tmp]"\n"
}' old_file > new_file

Another awk solution:

Code:

$ awk -F"," '{ if(NR>1){print $1;split($1,a," ");i=2;{while(a[3]-1!=0) {print a[1],a[2],NF-1,$(i);--a[3];++i}}}else print $0}' oldfile > newfile

Output from above:

Code:

ruby -ane 'BEGIN{a=[]};b=$F[3].split(",") if /^\d_q/;b.map!{|x|x="#{$F[0..2].join(" ")} #{x}"} if b; puts b if b ' file
1_q OMA1 2 43210
1_q OMA1 2 243007
1_q TFR1 3 429310
1_q TFR1 3 2412007
1_q TFR1 3 23366
1_q FTOMA1 5 3290
1_q FTOMA1 5 243317
1_q FTOMA1 5 24333
1_q FTOMA1 5 2433136
1_q FTOMA1 5 24334

Hi,

Thank you for your reply! When I use the awk code, I got the new file like below which contains additonal rows which I don't want. How to remove them? by the way, what does '/^\s*$/' mean? I don't understand. Thanks!

1_q OMA1 2 43210
1_q OMA1 2 243007
1_q OMA1 2
1_q TFR1 3 429310
1_q TFR1 3 2412007
1_q TFR1 3 23366
1_q TFR1 3
1_q FTOMA1 5 3290
1_q FTOMA1 5 243317
1_q FTOMA1 5 24333
1_q FTOMA1 5 2433136
1_q FTOMA1 5 24334
1_q FTOMA1 5

You get the additional rows when you skip the '/^\s*$/' line.

The line /^\s*$/ means any line that just contains white spaces or empty

If you notice your input file the format is as follows
1_q OMA1 2 43210,243007,

when a split is done on the above line using split($4,arr,","); the array arr has the values (43210,243007,'') . hence you get the additional line. by adding the check for empty values we can eliminate that row.

Hope this helps

If the only commas in the input occur in that last field:

Code:

tr -s , \\n < file | awk 'NF>1 {p=$0} NF==1 {f=$0; $0=p; $NF=f} 1'

f=final field's value
p=prefix of records with common origin


Regards,
Alister