We have used solid-state O-17 NMR experiments to determine the O-17 quadrupole coupling (QC) tensor and chemical shift (CS) tensor for the carbonyl oxygen in p-nitro-[1-O-17]benzaldehyde. Analyses of solid-state O-17 NMR spectra obtained at 11.75 and 21.15 T under both magic-angle spinning (MAS) and stationary conditions yield the magnitude and relative orientation of these two tensors: C-Q 10.7 +/-0.2 MHz, eta(Q) = 0.45 +/- 0.10, delta(11) = 1050 +/- 10, delta(22) = 620 +/- 10, delta(33) = -35 +/- 10, alpha = 90 +/- 10, beta = 90 +/- 2, gamma = 90 +/- 10 degrees. The principal component of the O-17 CS tensor with the most shielding, delta(33), is perpendicular to the H-C=O plane, and the tensor component with the least shielding, delta(11), lies along the C=O bond. For the O-17 QC tensor, the largest (chi(zz)) and smallest (chi(xx)) components are both in the H-C=O plane being perpendicular and parallel to the C=O bond, respectively. This study represents the first time that these two fundamental O-17 NMR tensors have been simultaneously determined for the carbonyl oxygen of an aldehyde functional group by solid-state O-17 NMR. The reported experimental solid-state O-17 NMR results provide the first set of reliable data to allow evaluation of the effect of electron correlation on individual CS tensor components. We found that the electron correlation effect exhibits significant influence on O-17 chemical shielding in directions within the H-C=O plane. We have also carefully re-examined the existing experimental data on the O-17 spin-rotation tensor for formaldehyde and proposed a new set of best "experimental" O-17 chemical shielding tensor components: sigma(11) = -1139 +/- 80, sigma(22) = -533 +/- 80, sigma(33) = 431 +/- 5, and sigma(iso) = -414 +/- 60 ppm. Using this new set of data, we have evaluated the accuracy of quantum chemical calculations of the 170 CS tensors for formaldehyde at the Hartree-Fock (HF), density-functional theory (DFT), Moller-Plesset second-order perturbation (MP2), and coupled-cluster singles and doubles (CCSD) levels of theory. The conclusion is that, while results from HF and DFT tend to underestimate the electron correlation effect, the MP2 method overestimates its contribution. The CCSD results are in good agreement with the experimental data.