Advanced use of awk

Question

Say you have a file like this:

Rowname  
useless number  
useless number  
1596 1787  
1595 1787  
1594 1787  
1565 1803  
1565 1804  

I know how to tell awk to skip the useless number and awk row 5 and 6 based on Rowname. BUT how can I tell awk to skip the row if the difference is only 1 from the previous row? The output should be:

1596 1787  
1565 1803

My problem is that in some cases I get this:

1595 1787  
1594 1787

Said another way: I always want the 3rd row from the row name, but I don't want row 4 (or 5 or 6) if the difference from row 3 only are 1 (or 2 or 3).

Does that make sense?

Answer 1

This seems to meet your requirement if you want the 'Rowname' printed; if not, omit the print in the first line of the script:

$     awk '/^Rowname/         { nr=NR; print; next }
>          NR==nr+1||NR==nr+2 { next }
>          NR==nr+3           { v1=$1; print; next }
>                             { if (abs(v1-$1) > 1) print; v1=$1; }
>          function abs(x)    { return (x < 0) ? -x : x }
>         ' data
Rowname
1596 1787
1565 1803
$

Answer 2

This may do it

awk '#print line 4
     NR == 4;
     #print any line that meets the "sum" comparison condition
     NR > 4 && ($1 + $2 - prev_sum != -1) && ($1 + $2 - prev_sum != 1);
     #set the sum for this line into a variable
    {prev_sum=$1 + $2}' file.txt