First, I would check that you are able to read the data as you expect. Most experienced users on this forum will tell you to avoid the Express VIs (the big blue boxes) and dynamic data (such as the "signals" output of Read from Measurement File). If your data is in fact a text file (.csv), use "Read Delimited Spreadsheet" instead. Connect the All Rows output from that function to a waveform graph; do you see your data displayed as you expect?
Once you know you're reading the data properly, you can feed it to the PID. You will want to put the PID in a loop, executing once for each row of data. You'll also want to wire the "dt" input of the PID for the actual time between measurements, since you'll be able to run it much faster than real time with data from a file.
Also, a note about terminology: the question in your post title is basically meaningless. You already have the process variable data - it's in your file! So no need to measure it, you've already done that, and no need for the PID. I'm assuming you actually want to measure the PID output.
