Sparse arrays have been recently applied to spectral sensing of wideband spectrum where the array is configured to estimate the directions of co-frequency emitters and also reconfigured, through antenna switching, for beamforming. The latter archives source signal isolation necessary for emitter identifications and signal modulation classifications and identifications. This paper examines the roles of the sparse array configuration and the sparse array weights in delivering the optimum and semi-optimum performance when adopting the criterion of maximum signal-to-interference and noise ratio (MaxSINR). Towards this end, we compare the source isolation performance based on the optimum sparse array configuration and array weights applied individually, jointly, or in sequence. The individually optimized sparse array configuration emerges when the array strives to orthogonalize the desired source steering vector and the interference subspace. Such sparse array can be followed by applying conventional beamformer weights or weights stemming from high interference to noise ratio assumption. HER curves are used in the comparison to delineate the offering of each array design approach.