Determining length of numerical ranges in file by perl/awk

Hi all,

I am now learning awk but get stuck on the following problem.

I have a file (input.txt) with various numerical ranges (before it, there is an id and :) in each line such as:

NP_416485.4: 4-5, 113-114, 395-399, 657-666, 671-675, 844-880, 889-889, 891-895, 963-966, 970-970, 991-992, 1126-1235
NP_417679.2: 309-413, 418-421, 441-442, 444-445, 447-481, 939-941, 943-984
NP_418770.2: 264-265, 267-272, 276-277, 287-288
NP_415931.4: 32-33, 73-75, 387-388, 394-396, 531-634

Now I want to go though each id and print only those ranges which are >=30. For the above input, the output should be:

NP_417679.2: 309-413, 447-481, 943-984
NP_415931.4: 531-634

I have tried this awk script:

awk -F"-" "{if(($2-$1)>=30) print $_}" input.txt

But it prints all the input ranges. Please suggest.

ps. I can also use any perl alternative

awk perl

• 2,375 views • link

updated 9.8 years ago by Jorge Amigo

• written 9.8 years ago by User 6777 •

2 answers

If you pre-process your input to add a trailing comma, you can use the space as a delimiter:

$ awk -F' ' '{ $0 = $0","; h = $1; f = ""; for (i = 2; i <= NF; i++) { r = substr($i, 0, length($i) - 1); split(r, re, "-"); if (re[2] - re[1] >= 30) { f = f" "$i; } } if (length(f) > 0) { print h" "substr(f, 2, length(f) - 2); } }' input.txt
NP_416485.4: 844-880, 1126-1235
NP_417679.2: 309-413, 447-481, 943-984
NP_415931.4: 531-634

• 0 views • link

written 9.8 years ago by Alex Reynolds

a quick perl alternative:

perl -ane '$s = "";
foreach (@F) { /(\d+)-(\d+)/ and $2 - $1 >= 30 and $s .= " $1-$2" }
print "$F[0]$s\n" if $s' input.txt

the -a option splits input by \s, it loops through all fields looking for valid intervals and appending them to an $s variable, which is printed at the end if not empty.

you can play around with the loop rationale in mind:

perl -ne '$s = ""; /^(\S+:)/ and $S = $1;
while (/(\d+)-(\d+)/g) { $s .= " $1-$2" if $2 - $1 >= 30 }
print "$S$s\n" if $s' input.txt

• 0 views • link

written 9.8 years ago by Jorge Amigo

sorry for my ignorance, is this a perl one liner? I'm on a windows machine and it shows error like:

>perl -ane '$s = "";
Can't find string terminator "'" anywhere before EOF at -e line 1.

C:\Users\cdutta\Desktop\x\New folder>foreach (@F) { /(\d+)-(\d+)/ and $2 - $1 >=
 30 and $s .= " $1-$2" }
'foreach' is not recognized as an internal or external command,
operable program or batch file.

• 0 views • link

written 9.8 years ago by User 6777 •

yes. I'm not sure how command line perl works on Windows, but you could try wrapping it all into a single command line before pasting it into the Windows console.

perl -ane '$s = ""; foreach (@F) { /(\d+)-(\d+)/ and $2 - $1 >= 30 and $s .= " $1-$2" } print "$F[0]$s\n" if $s' input.txt

• 0 views • link

written 9.8 years ago by Jorge Amigo

aaaam.. currently i'm trying it sir.. but certainly i'm missing something

• 0 views • link

written 9.8 years ago by User 6777 •

thanks for your answer

• 0 views • link

written 9.8 years ago by User 6777 •

this error:

Can't find string terminator '"' anywhere before EOF at -e line 1.

• 0 views • link

written 9.8 years ago by User 6777 •

as I say, I don't know how command line perl works on Windows. you could try to save the code as splitter.pl and then run it perl splitter.pl

• 0 views • link

written 9.8 years ago by Jorge Amigo