How can I extract a value from comma separated values in Perl?

Question

I have a log file containing statistics from different servers. I am separating the statistics from this log file using regex only. I am trying to capture the CPU usage from the running process. For SunOS, I have below output:

process,10050,user1,218,59,0,1271M,1260M,sleep,58.9H,0.02%,java

Here the CPU % is at 11th field if we separate by commas (,). This field has % sign which is unique and I can use below regex To get this value:

regex => q/^process,(?:.*?),((?:\d+)\.(?:\d+))%,java$/,

For the linux system I have below output:

process,26190,user1,20,0,1236m,43m,6436,S,0.0,1.1,0:00.00,java,

Here the CPU usage is at 10th column but without % sign and there is nothing unique I can see with this field.

What regex pattern should I use to get this value?

Answer 1

If the line is already comma separated, you can just use split on the string and pick the correct field.

e.g.

my @fields = split(/,/, $input);

Answer 2

you have data structure that has distinct delimiter, so don't use regex but just use splitting and get your item by index(or slicing). Its easier.

$output="process,10050,user1,218,59,0,1271M,1260M,sleep,58.9H,0.02%,java";
@s = split /,/,$output;
print "$s[10]\n";

for linux, just get $s[9]

Answer 3

I know nothing about Linux but just ignore the code if it looks too naive-minded :)

 /^process.*(?<=[A-Z],)((?:\d+)\.(?:\d+)).*java$/;

Answer 4

Use Text::CSV_XS to work with comma-separated values. If you're asking the question, you shouldn't be trying to handle it yourself. The module is extremely optimized and you won't be able to do a better job on your own.

Once you extract the right position, you can strip off the % sign if it's there.

Answer 5

Do you need to recognize the line as well, or only extract the value, i.e. do you expect some unrelated lines? If not the answer is /^(?:[^,]+,){9}([^,]+)/' this will extract the tenth field from some comma separated list