Smallest learning curve language to work with CSV files
Python is an excellent choice. The csv module makes reading and writing CSV files easy (even Microsoft's, uh, "idiosyncratic" version) and Python syntax is a breeze to pick up.
I'd actually recommend against Perl, if you're coming to it fresh. While Perl is certainly powerful and fast, it's often cryptic to the point of incomprehensible to the uninitiated.
What kind of calculation you have to do? Maybe R would be an alternative?
EDIT: just to give a few basic examples
# Basic usagedata <- read.csv("myfile.csv")# Pipe-separated valuesdata <- read.csv("myfile.csv", sep="|")# File with header (columns will be named as header) data <- read.csv("myfile.csv", header=TRUE)# Skip the first 5 lines of the filedata <- read.csv("myfile.csv", skip=5)# Read only 100 linesdata <- read.csv("myfile.csv", nrows=100)