awk double id sorter_filter

# This shell will get you the top x items for each group
# pe: 
#   if you have a csv with user_id, song_id, relevance
#   this will sort the csv by user_id asc and then by relevance desc
#   after that it will only print the top x results for each user to STDOUT

sort -t, -k 1,1n -k 3,3nr my.csv | awk -F, -v max_items=5 '{if (user_id != $1) {user_id=$1; count=1} ; if (count <= max_items) { print; } ; count+=1;}' > result.csv