bash: all combinations of lines

Question

I have the following file (this is semicolon delimited; the real file is tab-delimited)

abc;173959;172730
def;4186657;4187943
ghi;4703911;4702577
jkl;2243551;2242259

and I want to combine each line with each, so that my output would be:

abc;173959;172730;def;4186657;4187943
abc;173959;172730;ghi;4703911;4702577
abc;173959;172730;jkl;2243551;2242259
def;4186657;4187943;ghi;4703911;4702577
def;4186657;4187943;jkl;2243551;2242259
ghi;4703911;4702577;jkl;2243551;2242259

The order is not important.

I came up with the following awk-solution:

awk '{ a[$0] } END { for (i in a){ for (j in a){if (i != j)  print (i "\t" j) } } }' file

But this prints me the combinations in both directions, so for example

abc;173959;172730;def;4186657;4187943
def;4186657;4187943;abc;173959;172730

Because I am pretty unfamiliar with python or perl, I kindly ask for a solution using awk/bash etc.

Answer 1

In awk:

$ awk '{ a[$0] }
END {
    for(i in a) {
        delete a[i]            # new place for delete
        for(j in a)
            if(i!=j)
                print i ";" j
        # delete a[i]          # previous and maybe wrong place
    }
}' file
def;4186657;4187943;ghi;4703911;4702577
def;4186657;4187943;abc;173959;172730
def;4186657;4187943;jkl;2243551;2242259
ghi;4703911;4702577;abc;173959;172730
ghi;4703911;4702577;jkl;2243551;2242259
abc;173959;172730;jkl;2243551;2242259

Unfortunately the order is random.

Another way that restores the order and doesn't modify the a while processing (see comments) is:

$ awk '{ a[NR]=$0 }                  # index on NR
    END {
        for(i=1;i<=NR;i++)
            for(j=i+1;j<=NR;j++)     # j=i+1 is the magic
                print a[i] ";" a[j]
}' file
abc;173959;172730;def;4186657;4187943
abc;173959;172730;ghi;4703911;4702577
abc;173959;172730;jkl;2243551;2242259
def;4186657;4187943;ghi;4703911;4702577
def;4186657;4187943;jkl;2243551;2242259
ghi;4703911;4702577;jkl;2243551;2242259

Answer 2

This awk should work as well:

awk -F ';' 'NR==FNR{a[++k]=$0; next} {for (i=FNR+1; i<=k; i++) print $0 FS a[i]}' file{,}

abc;173959;172730;def;4186657;4187943
abc;173959;172730;ghi;4703911;4702577
abc;173959;172730;jkl;2243551;2242259
def;4186657;4187943;ghi;4703911;4702577
def;4186657;4187943;jkl;2243551;2242259
ghi;4703911;4702577;jkl;2243551;2242259

Answer 3

Could you please try following one, it will give you same order as Input_file's field values only, by reading the Input_file once only.

awk '{a[FNR]=$0} END{j=1;while(length(a)>=++k){for(q=j+1;q<=FNR;q++){print a[j]";"a[q]}j++};}'  Input_file

OR

awk  '
{
  a[FNR]=$0
}
END{
  j=1;
  while(length(a)>=++k){
     for(q=j+1;q<=FNR;q++){
       print a[j]";"a[q]
}
  j++
}
}
'   Input_file

Output will be as follows.

abc;173959;172730;def;4186657;4187943
abc;173959;172730;ghi;4703911;4702577
abc;173959;172730;jkl;2243551;2242259
def;4186657;4187943;ghi;4703911;4702577
def;4186657;4187943;jkl;2243551;2242259
ghi;4703911;4702577;jkl;2243551;2242259