Shell Programming and Scripting

NR == FNR is a common AWK idiom: an expression that returns true while reading the first input file (while the total number of input records seen so far is equal to the input record number in the current input file).

Excerption from the awk manual (nawk):

Regards

Does the above code work with all the records matched or unmatched?

lets say if i have an example like

file1

1 1 111 222 I 333 101 60 2 1 0 0 0 0 S 1 01 0 60 34340
1 1 111 222 I 333 101 60 3 1 0 0 0 0 S 1 01 0 60 34340
1 1 111 222 U 121 0 65 4 1 0 0 0 0 S 1 01 0 65 34340
1 1 111 222 U 231 0 125 5 1 0 0 0 0 S 1 01 0 125 34340
1 1 111 222 U 234 0 25 6 1 0 0 0 0 S 1 01 0 25 34340
1 1 123 212 I 342 336 150 2 1 0 0 0 0 S 1 01 0 150 34340
1 1 123 212 I 342 336 150 3 1 0 0 0 0 S 1 01 0 150 34340
1 1 123 212 U 421 0 65 4 1 0 0 0 0 S 1 01 0 65 34340
1 1 123 212 U 564 0 125 5 1 0 0 0 0 S 1 01 0 125 34340
1 1 458 878 U 636 0 225 1 1 0 0 Y 0 Y 0 0 0 0 0


file2

1 bb04 1 5153 123 6704 1310 111 SA 423386 0 0000 0000 49.00 14.00 34.00 0.00 00 4382151906251518 0000000000 00000 100091667 392071011 0000000000 0.00 0.00 0.00
1 823b 1 5153 125356926 6704 1615 392085746 SA 458313 0 0000 0000 279.00 34.00 244.00 0.00 00 0000000000 00000 100091684 392091255 0000000000 0.00 0.00 0.00



Here in this case
file1.col3 = file2.col8 and
file1.col4 = file2.col5

so if we find a match, then i would like to get all the fields of file1 along with file2.col7 and file2.col10 into file3

and if we dont find a match, can we output those records to a new file4

Code:

awk -F'\t' 'NR == FNR { _[$8,$5] = $7 FS $10; next }
($3,$4) in _ { print $0 FS _[$3,$4] > "file3"; next }
{ print > "file4" }' file2 file1

radoulov and frankiln....thanks for your answers.