Fred Otten came up with the following script using good old sed, awk and some plumbing, that draws up a nice list based on an input Lyx file.


#!/bin/sh
cat $1 | grep key\ \" |\
awk '{ print substr($2,2,length($2)-2)}' | \
sed -e s/,/\\n/g | \
awk 'BEGIN {i=0} \
{ if (temp[$1]) { temp[$1]=temp[$1]+1 } \
else { temp[$1]=1; tmp[i]=$1; i++; }; } \
END { for (j=0; j { print tmp[j] " " temp[tmp[j]] } }' | sort


This gives a two column listing of the citation keys and their frequency count. This of course can be extended using further awk statements to transpose the columns, or sort by frequency, rather than citation key.