I think it totally depends on the project and how the PI likes to run the projects. My PhD is in molecular biology and I have been working on all three data chapters simultaneously over the last 2.5 years and only in 1 is the data collection near completion. For the other two I may as well be back at square one...
Most people in my lab are set up with projects like this, but some have sequential projects whereby the success of one depends on the completion of another, and in that case I think one data chapter should aim to be completed each year, with emphasis on the word 'aim'.