CODEWITHSUNDEEP
bashawksedgrepsh
Shred
Shred

Reputation: 358

Reorder text file lines by a specific scheme with bash or awk

I need to figure how to reorder in bash a text file , made with another script, composed by lots of lines with a specific scheme. Here's 12 of them.

>NODE_3
nucleotide_cov: 170.3683
GC_CONT: 37.00
>NODE_18
nucleotide_cov: 168.8670
GC_CONT: 37.00
>NODE_23
nucleotide_cov: 178.0648
GC_CONT: 35.00
>NODE_41
nucleotide_cov: 174.4054
GC_CONT: 36.00

The output needed is:

GC_CONT: 37.00  nucleotide_cov: 170.3683    >NODE_3
GC_CONT: 37.00  nucleotide_cov: 168.8670    >NODE_18
GC_CONT: 35.00  nucleotide_cov: 178.0648    >NODE_23
GC_CONT: 36.00  nucleotide_cov: 174.4054    >NODE_41

where every column is divided by tab, so "\t" and GC_CONT needs to be the first value of them. awk solutions preferred.

EDIT

I'll try to more clear. Here's the output file by using 

awk 'NR%3{printf "%s ",$0;next;}1' input.txt 

>NODE_3 nucleotide_cov: 170.3683 GC_CONT: 37.00
>NODE_18 nucleotide_cov: 168.8670 GC_CONT: 37.00
>NODE_23 nucleotide_cov: 178.0648 GC_CONT: 35.00
>NODE_41 nucleotide_cov: 174.4054 GC_CONT: 36.00

Good, but i need to format them in order to have "GC_CONT:" at the beginning of every line.

Upvotes: 1

Views: 986

Answers (4)

potong
potong

Reputation: 58488

This might work for you (GNU sed):

sed -r ':a;N;/\n>/!s/(.*)\n(.*)/\2\t\1/;ta;P;D' file

Collect the lines needed for a record, swapping the lines and replacing the newlines by tabs.

Upvotes: 1

Nahuel Fouilleul
Nahuel Fouilleul

Reputation: 19325

Just for information as tagged sed :

sed -r '/^>/{N;N;s/(.*)\n(.*)\n(.*)/\3\t\2\t\1/g}'

Upvotes: 2

RomanPerekhrest
RomanPerekhrest

Reputation: 92884

Chhose one you like:

simple awk one-liner:

awk '/^>/{ n=NR; r=$0; next }{ r=$0 OFS r; if (NR-n==2) print r }' OFS='\t' input.txt

Or awk solution for strict order of lines:

awk '/^>/{ r1=$0; n=NR }
     n{ if (NR == n+1) r2=$0; else if (NR == n+2) print $0, r2, r1 }' OFS='\t' input.txt

The output:

GC_CONT: 37.00  nucleotide_cov: 170.3683    >NODE_3
GC_CONT: 37.00  nucleotide_cov: 168.8670    >NODE_18
GC_CONT: 35.00  nucleotide_cov: 178.0648    >NODE_23
GC_CONT: 36.00  nucleotide_cov: 174.4054    >NODE_41

Upvotes: 2

Pierre François
Pierre François

Reputation: 6073

Try this awk script:

/^>/    {node=$0}
/^nucl/ {nucl=$0}
/^GC/   {print $0 "\t" nucl "\t" node}

Or, from command line:

awk '/^>/{node=$0} /^nucl/{nucl=$0} /^GC/{print $0 "\t" nucl "\t" node}' input_file

Upvotes: 2

Related Questions