Diet and animal characteristics have a significant impact on the nitrogen (N)-use efficiency of dairy cows. A dataset (CowNflow) was built that compiles 28 N-balance experiments with Holstein dairy cows from 1983 to 2019, corresponding to 414 individual N flows, for a wide range of diets and animal characteristics. The dataset is composed of six Microsoft® Excel files that correspond to six levels of information. The main file, "CowNflow_6_Cow_measurements" reports individual weekly measurements of dry matter intake, daily faeces and urine excretion, milk production and composition, cow characteristics, and chemical composition of diets, faeces, urine, and milk. These raw data were used to calculate the N-balance, N-use efficiency, and nutrients' in vivo total-tract digestibility. The experiments, conducted under standardised conditions, had multiple aims and offered a wide range of diets. Consequently, each diet is classified according to the main forage offered, resulting in six diet types: (1) maize forage (maize silage or dehydrated maize) alone, (2) maize forage and dehydrated lucerne, (3) maize forage and grass hay, (4) maize forage and freshly cut herbage, (5) freshly cut herbage alone, and (6) dehydrated herbage. The other five Excel files provide supplementary information at larger scales and describe experiment characteristics, experimental treatments, offered feeds along with their chemical composition, ingredient composition of compound feeds, and cow characteristics. This dataset can be used to better understand animal and dietary determinants of N-use efficiency and the origin of N losses to the environment, to identify feeding strategies that reduce protein-rich concentrate use, and to decrease environmental impacts of dairy farming with a variety of foraging systems.