Joining Two Files Line by Line with awk

Suppose two files need to be joined line by line.

File a contains:

a 1
b 2
c 3

File b contains:

a 5
b 6

The desired result is:

a 1 5
b 2 6
c 3

The awk command is:

awk '{a[$1]=a[$1]" "$2}END{for (j in a) print j""a[j]}' a b

Honestly, this line is difficult to understand. Even after reading the awk syntax, I had not encountered this particular pattern. A Google search led to an explanation:

{ a[$1]=a[$1]" " $2; next } appends the second field ($2) to an array (a) indexed by the first field ($1).

In other words, the array a uses $1 as its key. The initial value of a[$1] is an empty string, and each matching " "$2 value is appended to it.

For example:

awk '{a[$1]=a[$1]"  "$2}END{for (j in a) print a[j]}' a b

The result, with two spaces between values, is:

  1  5
  2  6
  3

Another approach is:

paste -d' ' a b | awk '{print $1, $2, $4}'

The first solution is better because it still joins the records correctly when the first column in file b is in a different order. For example, the result remains the same when b contains:

b 6
a 5

Originally published on SegmentFault.