Under ideal conditions, Earth-based telescopes can observe near-Earth objects (NEOs) continuously from a few days to months during each apparition. Due to the usually complicated dynamics of the Sun-Earth-NEO triplet, the time interval between consecutive apparitions typically ranges from months to several years. On these time scales, exiguous single-apparition sets (SASs) of observations having short observational time-intervals lead to substantial orbital uncertainties. Linking of SASs over apparitions thus becomes a nontrivial task. For example, of a total of roughly 4,100 NEO observation sets, or orbits, currently known, some 2,300 are SASs, for which the observational time interval is less than 180 days. Either these SASs have not been observed at an apparition following the discovery apparition or the linkage of SASs has failed, an option which should preferably be eliminated. As a continuation to our work on the short-arc linking problem at the discovery moment (Granvik & Muinonen, 2005, Icarus 179, 109), we have investigated the possibility of using a similar method for linking exiguous SASs over apparitions. Assuming that the observational time-interval for SASs of NEOs is typically at least one day (minimum requirement set by the Minor Planet Center), the orbital-element probability-density function is constrained as compared to the typical short-arc case with an observational time interval of only a few tens of minutes. Because of the smaller orbital-element uncertainty, we can use the short-arc method (comparison in ephemeris space) for longer time spans, or even do the comparison directly in the orbital-element space (Cartesian, Keplerian, equinoctial, etc.), thus allowing us to assess the problem of linking SASs of NEOs. Due to possible close approaches with the Earth and other planets, and substantial propagation intervals, we have developed new n-body techniques for the orbit computation.