We construct the equivalent width distribution of Ca II K absorption lines seen in lines-of-sight through the galactic halo from an unbiased sample of extragalactic sightlines, and compare the absorption by the interstellar medium of our own Galaxy with that seen in the spectra of QSOs lying behind low-redshift galaxies. The Ca II absorption seen in several QSO spectra apparently arises in material deposited inhomogeneously around the foreground galaxy, perhaps as a result of earlier galaxy interactions or mergers. We demonstrate that for other QSO sightlines such as those toward 0446-208 and 2020-370, the measured column densities of the absorbing gas can be reproduced by assuming that the quasar line-of-sight passes through the inclined disc of the intervening galaxy. In contrast, we find that such a model has difficulty in explaining the non-detection of Ca II in the pair IC 1746-PHL 1226, and requires the diameter of the galaxy in the foreground of the quasar 1543+489 to be >= 105 h^-1^ kpc if responsible for the absorption. Since Wlambda_(H) in the Galaxy is correlated with the number of components comprising the line, the large values of W_lambda_(K) seen in most QSO-galaxy pairs can similarly be accounted for by assuming that a line is composed of many individual absorbers spread out over larger velocities than those found locally. This explanation is consistent with absorption attributed to a sightline through an inclined disc, where the circular rotation of clouds will produce absorption over a large velocity range, and with absorption from material deposited after galaxy interactions. Finally, the poor success in detecting Ca II in the haloes of galaxies at the present epoch does not necessarily rule out the future detection of Mg II along the same lines-of-sight. At z~0.5, there are still insufficient data to decide if the Ca II absorption found in Mg II systems arises in the same absorbing material as the Mg II.