Display message when no match found in AWK - bash

Question

I'm writing a small BASH script that reads a csv file with names on it and prompts the user for a name to be removed. The csv file looks like this:
Smith,John
Jackie,Jackson
The first and last name of the person to be removed from the list are saved in the bash variables $first_name and $last_name.
This is what I have so far:
cat file.csv | awk -F',' -v last="$last_name" -v first="$first_name" ' ($1 != last || $2 != first) { print } ' > tmpfile1
This works fine. However, it still outputs to tmpfile1 even if no employee matches that name. What I would like is to have something like:
if ($1 != last || $2 != first) { print } > tmpfile1 ; else { print "No Match Found." }
I'm new to awk and can't get that last part to work.
NOTE: I do not want to use something like grep -v "$last_name,$first_name"; I want to use a filtering function.

Answer 1

You can redirect right inside the awk script, and only output matches found.
awk -F',' -v last="$last_name" -v first="$first_name" '
$1==last && $2==first {next}
{print > "tmpfile"}
' file.csv
Here are some differences between your script and this....
This has awk reading your CSV directly, rather than having UUOC.
This actively skips the records you want to skip,
and prints everything else through a redirect.
Note that you could, if you wanted, specify the target to which to redirect in a variable you pass in using -v as well.
If you really want the "No match found" error, you can set a flag, then use the END special condition in awk...
awk -F',' -v last="$last_name" -v first="$first_name" '
$1==last && $2==first { found=1; next }
{ print > "tmpfile" }
END { if (!found) print "No match found." > "/dev/stderr" }
' file.csv
And if you want no tmpfile to be created if a match wasn't found, you would either need to scan the file TWICE, once to verify that there's a match, and once to print, or if there's no risk that the size of the file would be too great for available memory, you could keep a buffer:
awk -F',' -v last="$last_name" -v first="$first_name" '
$1==last && $2==first { next }
{ output = (output ? output ORS : "" ) $0 }
END {
if (output)
print output > "tmpfile"
else
print "No match found." > "/dev/stderr"
}
' file.csv
Disclaimer: I haven't tested any of these. :)

Answer 2

You can do two passes over the file, or you can queue up all of the file so far in memory and then just fail if you reach the END block with no match.
awk -v first="$first" last="$last" '$1 != last || $2 != first {
for (i=1; i<=n; ++i) print a[i] >>"tempfile"; p=1; split("", a); }
# No match yet, remember this line for later
!p { a[++n] = $0; next }
# If we get through to here, there was a match
p { print >>"tempfile" }
END { if (!p) { print "no match" >"/dev/stderr"; exit 1 } }' filename
This requires you to have enough memory to store the entire file (this will be required when there is no match).

Answer 3

With a bash script, you can test if awk print something.
If yes, remove the tmpfile.
c=$(awk -F',' -v a="$last_name" -v b="$first_name" '
$1==a && $2==b {c=1;next}
{print > "tmpfile"}
END{if (!c){print "no match"}}' infile)
[ -n "$c" ] && { echo "$c"; rm tmpfile;}