With the developments of modern space information sciences, specifically relying on the quickly updated and large scale remote sensing observation technology, the derivative LUC (land use and land cover) data generations, the strong spatial analysis functions and the visual information representation functions supported by geographic information system (GIS) technology, the traditional social and economic statistics data with Relational format can be transferred to the spatial data (Vector or Grid format). This paper firstly analyzed the shortcomings of traditional statistical data in the new-era applications, which characterized by long update-period, low spatial-resolution, poor visualization effects and non-support for the spatial computing and spatial analysis. We briefly reviewed the history of theoretical studies and database constructions of society and economy grid database. Based on the clarification about the concepts of grid process of society and economy statistics data, three basic requirements, i.e. temporal compatibility, space reference system compatibility and self-consistent logic, were put forward as the main principles in the transferring process, and a series of key index including 24 items to build a national scale society and economy grid database were posted. Then, based on the above data transferring principles and index system, we discussed the 5-step transferring processes which include (1) logical checking, (2) spatial matching, (3) code matching, (4) spatial discretization, and (5) checking and calibration. We also analyzed the general models to transfer the rational datasets into grid dataset in detail, and discussed the corresponding strategies when people integrate different datasets. Finally, we summarized the common questions and suggested that the spatial discrete model construction and the quick update routine works to prepare the background resources/environment datasets are the main focuses in the future studies.