Understanding the intricate synaptic connectivity in living neural circuits is crucial for unraveling the relationship between network structure and function, as well as its evolution during development, learning, and recovery from injury. However, current methodologies for identifying connected neurons in vivo suffer from limitations, particularly with regards to their throughput. In this study, we introduce a groundbreaking framework for in vivo connectivity mapping that combines two-photon holographic optogenetics for activating single or multiple potential presynaptic neurons, whole-cell recording of postsynaptic responses, and a compressive sensing strategy for efficiently retrieving individual postsynaptic neurons' responses when multiple potential presynaptic neurons are simultaneously activated. The approach was validated in the layer 2/3 of the visual cortex in anesthetized mice, enabling rapid probing of up to 100 cells in approximately 5 minutes. By identifying tens of synaptic pairs, including their connection strength, kinetics, and spatial distribution, this method showcases its potential to significantly advance circuit reconstruction in large neuronal networks with minimal invasiveness. Moreover, through simultaneous multi-cell stimulation and compressive sensing, we demonstrate up to a three-fold reduction in the number of required measurements to infer connectivity with limited loss in accuracy, thereby enabling high-throughput connectivity mapping in vivo. These results pave the way for a more efficient and rapid investigation of neuronal circuits, leading to deeper insights into brain function and plasticity.