Fixed effects for ID in a linear regression?

Hi all,

Just a rather simple question that I can't get my head around. I have a dataset in a long format with around 1000 IDs. About 200 of them occur two times (the collection of the data was over two years), while the rest are only included once. I want to make a linear regression (not a panel regression), but how should I control for the recurring IDs? Should I include fixed effect for ID, or is it better to include fixed effect for year?

Best,
Marcus