You don't need to know any of that. All you need to know is:
- You have columnar output with the same number of columns in each row.
- The first column is the user login name.
- The 4th column is the %MEM column.
- Awk initializes variable at 0 automatically.
Awk index keys are arbitrary values :)
ps aux | awk 'NR != 1 {x[$1] += $4} END{ for(z in x) {print z, x[z]"%"}}'
I pipe the input to awk, and then I tell it to skip line 1 (NR != 1). Then for every single line, it reads in English like this:
in array 'x', increment the value in
'x[username]' by the value in the 4th
column. If the array key doesn't exist
yet, create it, initialized at 0, and
then increment that by the value in
the 4th column.
When Awk has done this for every single line, it runs the code in the 'END' block, which says to print out each user name (z in x), and the final aggregated value associated with that user (x[z]), followed by a "%".
If you want to see it build line by line, you can use the same basic rules, but add a print after each line is processed instead of in the 'END':
ps aux | awk 'NR != 1 {x[$1] += $4; print $1, x[$1]}'