Use bash-foo to print all Apache log 404s in the order how often they appeared

evaluate_404s

#!/bin/bash

LOGFOLDER=/var/log/apache2
FIELD=7
HTTP_CODE=404
LOG_BASENAME=access.log

cd "$LOGFOLDER"

# print all zipped logfiles (suppressing errors)
zcat "$LOG_BASENAME".*.gz 2>/dev/null | \
# print current logfile plus the previously zipped ones
cat "$LOG_BASENAME" - | \
# search for matching errors
grep 'HTTP/1\.." '"$HTTP_CODE" | \
# print the field containing the requested URL
awk '{ print $'"$FIELD"'; }' | \
# prepend each line with how often it already occured
# @see <http://stackoverflow.com/a/15389597/113195>
awk '{if(order[$1]==0)order[$1]=++counter;print order[$1]" "$0;}' | \
# sort by this number
sort -n | \
# and reverse order
tac | \
# remove every line, whose URL had already appeared
awk '{if(order[$2]==0){order[$2]=1;print $0;}}' | \
# and sort again, so that the oftenmost URLs land at the bottom
sort -n

The response looks something like this:

17 /missing.image
22 /moved-article
293 /data:image/png;base64,iu85gjk856+bir8tgilt84fsad8962bvjhgas876hg
691 /wp-login