CODEWITHSUNDEEP
bashunixawk
efrem
efrem

Reputation: 455

Compare two columns in different files and append data for the shared items - UNIX

I have been searching in the forum a solution for my problem but unfortunately I was not successful. I am quite green in bash; therefore I would really appreciate any suggestion ot link to previous question/discussions

I have two files:

file1

    Ada     ada
    Ada     ada
    Ada     aidB
    Ada     alkA
    Ada     alkB
    Ada     alkB
    AdiY    adiA
    AdiY    gadA

file2

    NP_414557.1     insL1
    NP_414559.1     mokC
    YP_025292.1     hokC
    NP_414560.1     nhaA
    NP_414561.1     nhaR
    NP_414562.1     insB1
    NP_414563.1     insA
    NP_414564.1     rpsT

I would like to compare file1 and file2 based on the second column. If the object is the same I would like to append the text in the column 1 in file2 relative to that specific shared entry in a third column in file1.

Expected output: file1

    PhoB  caiT  NP_414580.1     
    PhoP  caiE  NP_414581.1     
    PhoP  caiB  NP_414582.1

Please consider the text reported just as example.

I am struggling with awk trying to adjust the code from a previouse question:

Compare two files of different columns and print different columns

awk 'NR==FNR{a[$1]=$2 ; next }$2 in a {$1=$1 FS a[$2]print}' file1 file2

Thank you very much in advance for your time and your help.

Best.

Upvotes: 3

Views: 5520

Answers (2)

Josh Jolly
Josh Jolly

Reputation: 11786

As @fedorqui commented, your example inputs/output are not consistent. I think this should do the trick though:

awk 'NR==FNR{a[$2]=$0; next} a[$2]>0{print a[$2],$1}' file1 file2

file1:

A alice
B bob
C carol
D dan

file2:

1 dan
2 alice
3 carol
4 bob

Output:

$ awk 'NR==FNR{a[$2]=$0} NR>FNR && a[$2]>0{print a[$2],$1}' file1 file2
D dan 1
A alice 2
C carol 3
B bob 4

Output can be sorted by whatever column you choose using sort -k. Breaking down the awk code:

  • NR==FNR{a[$2]=$0; next} - NR is an awk variable which contains the total number of lines processed so far. FNR is similar, but only contains the number of processes lines of the current file, so this condition effectively means "only do this for the first input file". The associated action stores the entire line ($0) in the associative array a, with the index being the value of the second field of the row. next just means that awk should move to the next line without doing further processing.

  • a[$2]>0{print a[$2],$1} - this condition will only be tested for the second and subsequent input files. The second field is used as an index to look up a value from the a - if the value is greater than 0 (ie, the value is a string in this case) then the value is printed, followed by the first field of the current line.

Basically, every line of the first file is stored in an array, indexed by the second field of the line. If that second field is matched in the second file, then the whole line from the first file is printed, with the second field from the second file appended.

Upvotes: 3

antoniochiurla
antoniochiurla

Reputation: 72

You can use join command to link fields on files. The simplest way to use it can be:

 join -j 2

to obtain an output with the first and second file of first file and following the first field of second file you can use the -o option as follow:

 join -j 2 -o 1.1 1.2 2.1

it is required that the lines in both files are sorted on second field, if the lines are not correctly sorted you can use the following bash trick:

 join -j 2  -o 1.1 1.2 2.1 <(sort -k2 1) <(sort -k2 2)

Supposing this is the content of two files:

 $ cat 1
 FIRST   first
 SECOND  second
 THIRD   third
 FOURTH  fourth
 $ cat 2
 2       second
 1       first
 3       third
 4       fourth

Note: the field separator on each line is the TAB character

The final result:

 $ join -j 2 -o 1.1 1.2 2.1 <(sort 1) <(sort -k2 2)
 FIRST first 1
 FOURTH fourth 4
 SECOND second 2
 THIRD third 3

Upvotes: 1

Related Questions