awk in while statement

Question

Eu tenho a seguinte entrada:

cat moldata
>species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT

>species_2
CCAAGGTCATCCATGACAACTTTGAGATCATTGAAGGCCTGATGACCACTGTACACGCCACCACCGCTACTCAGAAGA
GTCGACGGACCTTCCGGTAAACTCTGGCGTGATGGTCGTGGCGCTCAACAAAACATCATTCCCGCCTCTACTGGTGCT

>species_3
CAAAGCCGTAGGCAAAGTCATTCCTGCTCTCAACGGTAAACTGACTGGCATGGCCTTCCGTGTTCCCGTTCCAAATGT
CGGTTGTGGATCTTACTGTTCGCyTGGGAAAACCAGCCTCTTATGACrCCATTAAACAGAAGGTCAAGGAGGCTGCTG

>species_4
GGTCCTTTGAAGGGTATTCTTGGATACACCGAAGATCAAGTTGTGTCCACCGACTTTGTTGGAGACACACACTCTTCA
CTTTGACGCTGCTGCTGGTATCTCCCTCAACGATAACTTCGTCAAACTTATCAGCTGGTACGACAATGAATATGGATA

>species_5
GTTCCGCAAAGCTCAATGCCCTATTGTTGAGCGTCTGACCAATTCTCTCATGATGCATGGCCGCAACAACGGCAAGAA
TGATGGCAGTGCGAATTGTTAAGCATGCCTTTGAAATCATCCACCTTCTGACTGGAGAGAATCCTCTTCAAGTACTCG

e eu gostaria de recuperar da moldata o registro da espécie (nome da espécie + o bloco de linhas que a segue até o próximo registro de espécie) para as espécies listadas aqui:

cat species_list
species_1
species_3
species_5

para obter a seguinte saída:

cat output
>species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT

>species_3
CAAAGCCGTAGGCAAAGTCATTCCTGCTCTCAACGGTAAACTGACTGGCATGGCCTTCCGTGTTCCCGTTCCAAATGT
CGGTTGTGGATCTTACTGTTCGCyTGGGAAAACCAGCCTCTTATGACrCCATTAAACAGAAGGTCAAGGAGGCTGCTG

>species_5
GTTCCGCAAAGCTCAATGCCCTATTGTTGAGCGTCTGACCAATTCTCTCATGATGCATGGCCGCAACAACGGCAAGAA
TGATGGCAGTGCGAATTGTTAAGCATGCCTTTGAAATCATCCACCTTCTGACTGGAGAGAATCCTCTTCAAGTACTCG

Eu tentei fazer awk funcionar em um loop while:

while read line; 
do 
    if grep -q "$line" moldata; 
    then echo $line |  awk -v line=${line} 'BEGIN {RS=">"} /line/ {print $0}' moldata >> output; 
    else echo "$line not found"; 
    fi; 
done < species_list

Eu li sobre a opção getline com awk , mas não consegui fazer isso funcionar.

    

Answer 1

Se você insistir em fazer isso com awk :

( echo -n '/^>('; paste -sd\| - <species_list | tr -d '\n'; echo -n ')/,/^$/' ) | \
    awk -f - moldata

    

Answer 2

Para um único registro, você pode fazer algo assim:

awk '/species_1/{print;while (getline line){if(line !~/species/) print line; else break} }' input.txt                  
>species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT

Com vários itens, você pode querer algo ao longo destas linhas:

$ while IFS= read -r line                                                                                                
> do
>   awk -v spec="$line" '$0~spec{print;while (getline line){if(line !~/species/) print line; else break} }' input.txt    
> done < species.txt                                                                                                     
>species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT

>species_3
CAAAGCCGTAGGCAAAGTCATTCCTGCTCTCAACGGTAAACTGACTGGCATGGCCTTCCGTGTTCCCGTTCCAAATGT
CGGTTGTGGATCTTACTGTTCGCyTGGGAAAACCAGCCTCTTATGACrCCATTAAACAGAAGGTCAAGGAGGCTGCTG

>species_5
GTTCCGCAAAGCTCAATGCCCTATTGTTGAGCGTCTGACCAATTCTCTCATGATGCATGGCCGCAACAACGGCAAGAA
TGATGGCAGTGCGAATTGTTAAGCATGCCTTTGAAATCATCCACCTTCTGACTGGAGAGAATCCTCTTCAAGTACTCG

Solução alternativa usando dois arquivos como argumentos

Sabendo que os dados no exemplo do OP são sempre duas linhas após o nome da espécie, podemos carregar os nomes das espécies em species.txt no array e, em seguida, usar o loop para ler uma linha duas vezes e imprimi-la.

$ awk 'FNR==NR{species[$0]}; NR!=FNR{ sub(/>/,"");if ($0 in species){ print $0; for(i=0;i<=1;i++) {getline data;print data}  }}' species.txt input.txt
species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT
species_3
CAAAGCCGTAGGCAAAGTCATTCCTGCTCTCAACGGTAAACTGACTGGCATGGCCTTCCGTGTTCCCGTTCCAAATGT
CGGTTGTGGATCTTACTGTTCGCyTGGGAAAACCAGCCTCTTATGACrCCATTAAACAGAAGGTCAAGGAGGCTGCTG
species_5
GTTCCGCAAAGCTCAATGCCCTATTGTTGAGCGTCTGACCAATTCTCTCATGATGCATGGCCGCAACAACGGCAAGAA
TGATGGCAGTGCGAATTGTTAAGCATGCCTTTGAAATCATCCACCTTCTGACTGGAGAGAATCCTCTTCAAGTACTCG

    

Answer 3

declare -a list
readarray list < /path/to/species_list
for species in ${list[@]}; do
   sed -ne "/$species/,/^\$/p" moldata
done

Normalmente, é possível citar um script sed embutido, mas estamos usando uma variável shell para procurar seus dados e, portanto, ter que escapar do $ para evitar que o shell tente analisá-lo. / p>     

Answer 4

"Por uma dispensação especial, uma string vazia como o valor de RS indica que os registros são separados por uma ou mais linhas em branco", então

awk '    BEGIN { OFS=FS="\n"; RS=""; ORS="\n\n" }
         FNR==NR { for (k=0;k<NF;) ++species[">"$++k] }
         $1 in species
' species_list moldata

    

Answer 5

$ awk '
    NR==FNR{ sp[$0]; next }        # this is for reading species data into sp array
    /species/ {                    # from this point for moldata's lines
        if (substr($0,2) in sp) {  # if the read line match with data in sp array
            print                  # first print title
            while(getline > 0) {   # and then read until blank line and print
                if (NF == 0) { print ""; break }
                print 
            } 
        }
    }
' species_list moldata 
>species_1
?????????CACTTGGArGGTGGAGCCAAGAAGGTTATTATTTCTGCTCCCAGTGCTGACGCGCCCATGTACGTGGTC
TGTCAACCTCGATTCTTATGACCCATCTGCTAAGGTCATTTCGAATGCTTCCTGCACCACCAACTGCCTCGCTCCCCT

>species_3
CAAAGCCGTAGGCAAAGTCATTCCTGCTCTCAACGGTAAACTGACTGGCATGGCCTTCCGTGTTCCCGTTCCAAATGT
CGGTTGTGGATCTTACTGTTCGCyTGGGAAAACCAGCCTCTTATGACrCCATTAAACAGAAGGTCAAGGAGGCTGCTG

>species_5
GTTCCGCAAAGCTCAATGCCCTATTGTTGAGCGTCTGACCAATTCTCTCATGATGCATGGCCGCAACAACGGCAAGAA
TGATGGCAGTGCGAATTGTTAAGCATGCCTTTGAAATCATCCACCTTCTGACTGGAGAGAATCCTCTTCAAGTACTCG