同一行上的多个grep输出

Question

这似乎是一个非常微不足道的问题，但我没有足够的经验使用grep和echo来自己回答。我看了here和here没有成功。

我有一个文件，其开头像这样（.gff文件）超过1,000,000行。

NW_007577731.1 RefSeq region 1 3345205 . + . ID=id0;Dbxref=taxon:144197;Name=Unknown;chromosome=Unknown;collection-date=16-Aug-2005;country=USA: Emerald Reef%2C Florida;gbkey=Src;genome=genomic;isolate=25-593;lat-lon=25.6748 N 80.0982 W;mol_type=genomic DNA;sex=male 
NW_007577731.1 Gnomon gene 7982 24854 . - . ID=gene0;Dbxref=GeneID:103352799;Name=LOC103352799;gbkey=Gene;gene=LOC103352799;gene_biotype=protein_coding 
NW_007577731.1 Gnomon mRNA 7982 24854 . - . ID=rna0;Parent=gene0;Dbxref=GeneID:103352799,Genbank:XM_008279367.1;Name=XM_008279367.1;gbkey=mRNA;gene=LOC103352799;model_evidence=Supporting evidence includes similarity to: 22 Proteins%2C and 73%25 coverage of the annotated genomic feature by RNAseq alignments;product=homer protein homolog 3-like;transcript_id=XM_008279367.1 
NW_007577731.1 RefSeq region 1 3345205 . + . ID=id0;Dbxref=taxon:144197;Name=Unknown;chromosome=Unknown;collection-date=16-Aug-2005;country=USA: Emerald Reef%2C Florida;gbkey=Src;genome=genomic;isolate=25-593;lat-lon=25.6748 N 80.0982 W;mol_type=genomic DNA;sex=male 
NW_007577731.1 Gnomon gene 7982 24854 . - . ID=gene0;Dbxref=GeneID:103352799;Name=LOC103352799;gbkey=Gene;gene=LOC103352799;gene_biotype=protein_coding 
NW_007577731.1 Gnomon mRNA 7982 24854 . - . ID=rna0;Parent=gene0;Dbxref=GeneID:103352799,Genbank:XM_008279367.1;Name=XM_008279367.1;gbkey=mRNA;gene=LOC103352799;model_evidence=Supporting evidence includes similarity to: 22 Proteins%2C and 73%25 coverage of the annotated genomic feature by RNAseq alignments;product=homer protein homolog 3-like;transcript_id=XM_008279367.1 

我想到grep上在第三列中含有mRNA得到这个制表符分隔的输出（值的字段gene=，product=，transcript_id=）线。

LOC103352799 homer protein homolog 3-like XM_008279367.1 
LOC103352799 homer protein homolog 3-like XM_008279367.1 

随着离谱缺乏优美的，我可以分别使用

grep "mRNA\t" myfile.gff|sed s/gene=/@/|cut -f2 -d"@" |cut -f1 -d";" 
grep "mRNA\t" myfile.gff|sed s/product=/@/|cut -f2 -d"@" |cut -f1 -d";" 
grep "mRNA\t" myfile.gff|sed s/transcript_id=/@/|cut -f2 -d"@" |cut -f1 -d";" 

得到3列，但如何可以附加在同一行这3个命令的输出？我曾尝试

echo -e "`grep "mRNA\t" myfile.gff|sed s/gene=/@/|cut -f2 -d"@" |cut -f1 -d";"`\t`grep "mRNA\t" myfile.gff|sed s/product=/@/|cut -f2 -d"@" |cut -f1 -d";"`\t`grep "mRNA\t" myfile.gff|sed s/transcript_id=/@/|cut -f2 -d"@" |cut -f1 -d";"`" 

但这里是输出：

LOC103352799 
LOC103352799 homer protein homolog 3-like 
homer protein homolog 3-like XM_008279367.1 
XM_008279367.1 

非常感谢您的帮助！

Answer 1

用awk：

$ awk 'BEGIN { 
    FS=OFS="\t"      # field separators to tab 
    k="gene,product,transcript_id" # keyword list 
    split(k,a,",")     # split keywords to a hash for matching 
    for(i in a)      # values to keys 
     p[a[i]] 
} 
$3=="mRNA" { 
    b=""        # reset buffer b 
    split($9,a,"[=;]")    # split the data to a hash 
    for(i in a)      # iterate and search 
     if(a[i] in p)     # ... for keywords, if match, 
      b=b (b==""?"":OFS) a[i+1] # ... value is the next, buffer 
    print b       # output buffer 
}' file 
LOC103352799 homer protein homolog 3-like XM_008279367.1 
LOC103352799 homer protein homolog 3-like XM_008279367.1 

Answer 2

谈到oneliners，这里有一个在sed：

sed -nE '/\tmRNA\t/ { s/.*gene=([^;]+).*product=([^;]+).*transcript_id=([^;]+)/\1\t\2\t\3/g;p }' file 

唯一的假设是gene，product和transcript_id领域的固定顺序。这可以通过一些改变来解决，但是由于正则表达式的可读性。