将带有一列的 file1 与来自 file2 的两列进行比较

Question

I have two files: file1.txt and file2.txt

#file1.txt
xap1
NM_121
abc4
xxx0
uvw


#file2.txt
A123  001  xap1    mmmmm
B123       xxx0    nnnnn
C123  003  yyy1    ppppp
D123  004  zzz1    NM_121
E123  005  abc4    llllll
F123       jjjj    www

I want following output based on matching of column 1 of file1 with column 3 and column 4 of file2, get column 2 from file2 and print both:

#file3.txt
xap1    001
NM_121  004
abc4    005
xxx0    NA
uvw     NA

I used the following command, but don't know how to print column 1 from file1:

grep -w -F -f file1.txt file2.txt | awk '{print $2) > file3.txt

Thanks.

Answer 1

Could you please try following, this could be done in a single awk itself (written and tested with shown samples only).

awk '
FNR==NR{
  a[$3]=$2
  a[$4]=$2
  next
}
{
  printf("%s\n",$1 in a?$1 OFS a[$1]:$1 OFS "NA")
}
' Input_file2  Input_file1

One liner form as follows:

awk 'FNR==NR{a[$3]=$2;a[$4]=$2;next}{printf("%s\n",$1 in a?$1 OFS a[$1]:$1 OFS "NA")}' file2.txt file1.txt

Output will be as follows.

xap1 001
NM_121 004
abc4 005
xxx0 NA

---

---

OR: For windows systems, try following by making an .awk script(named script.txt ), here is a very nice link which can guide you more how to run awk on windows systems too Convert Text to Table (Space Delimited or Fixed length)

• If you have installed Windows Subsystem for Linux, you can directly execute the awk script as described above on the bash command line. If you have installed (or going to install) gawk as an independent application software, following guidance will help:

• First download Gawk for Windows from an appropriate server such as Sourceforge. There are two types of installation: with installer or without installer. The choice is up to you. Following description is based on the case without installer.

• Unzip the downloaded file to extract binaries and modules in an arbitrary location. (Download folder, desktop, or wherever).

• Create a working folder with an arbitrary name (such as "myawk") on your desktop or wherever convenient.

• Copy the script below to a file with an arbitrary name (such as "script.txt").

As awk executable doesn't care about the extension of the script file, you can keep it with ".txt" to associate a text editor or can change to ".awk" for specification.

FNR==NR{
  a[$3]=$2
  a[$4]=$2
  next
}
{
  printf("%s\n",$1 in a?$1 OFS a[$1]:$1 OFS "NA")
}

Now type following command on your terminal:

C:\your\path\to\gawk.exe -f script.txt Input_file2  Input_file1

Answer 2

$ cat tst.awk
BEGIN { FS=OFS="\t" }
NR==FNR {
    map[$3] = map[$4] = ($2 == "" ? "NA1" : $2)
    next
}
{ print $1, ($1 in map ? map[$1] : "NA2") }

$ awk -f tst.awk file2 file1
xap1    001
NM_121  004
abc4    005
xxx0    NA1

I used 2 different NA values to distinguish between cases where $1 exists in file 2 but has a blank entry (eg xxx0 ) vs where it just doesn't exist in file2, eg some random string like foobar :

$ cat file1
xap1
NM_121
abc4
xxx0
foobar

$ awk -f tst.awk file2 file1
xap1    001
NM_121  004
abc4    005
xxx0    NA1
foobar  NA2

Massage to suit.

Answer 3

The following:

{
join -t$'\t' -12 -23 -o1.1,1.2,2.2 <(nl -w1 file1.txt | sort -t$'\t' -k2) <(sort -t$'\t' -k3 file2.txt)
join -t$'\t' -12 -24 -o1.1,1.2,2.2 <(nl -w1 file1.txt | sort -t$'\t' -k2) <(sort -t$'\t' -k4 file2.txt)
} |
sort -t$'\t' -k1 | cut -f2- |
# insert NA is it's missing value
sed 's/\t$/\tNA/'

With the following recreation of input files:

cat <<EOF >file1.txt
xap1
NM_121
abc4
xxx0
EOF

# used tr to recreate a tab separated file
tr ' ' '\t' <<EOF >file2.txt
A123 001 xap1 mmmmm
B123  xxx0 nnnnn
C123 003 yyy1 ppppp
D123 004 zzz1 NM_121
E123 005 abc4 llllll
EOF

Outputs:

xap1    001
NM_121  004
abc4    005
xxx0    NA

Tested on repl .

Short explanation of main points:

• nl -w1 file1.txt | sort -t$'\\t' -k2 nl -w1 file1.txt | sort -t$'\\t' -k2 - number lines in file2.txt and sort with the second field
• join - join files. We join file1.txt with file2.txt twice - first on first field and 3rd field and then on first field and 4th field from file1.txt and file2.txt. For join inputs have to be sorted on the joined fields.
• sort -t$'\\t' -k1 | cut -f2- sort -t$'\\t' -k1 | cut -f2- - the lines in file1.txt are numbered, so later we can sort them using the line numbers (ie. restore original sorting order of file1.txt) and remove the line numbers
• sed 's/\\t$/\\tNA/' - the field is empty in file2.txt , while OP specified the output to be "NA". If the second column is missing from the output, insert the chracters NA there.
• >( ... ) is a process substition
• If a value is matched with both 3rd and 4rd field, it will be twice in the output. It could be just piped via sort -k1 -u to remove the duplicates, depending on needs.

The sorting and numbering of file1.txt could be optimized with some tee like nl | sort | tee >(join - file2.txt) >(join - file2.txt) nl | sort | tee >(join - file2.txt) >(join - file2.txt)