Hydrologists seek to better understand the processes involved in catchment response during floods using hydrological models. Model development requires databases combining hydrological, climatic and physical data. Large catchment databases have been developed at Irstea (Antony, France) over the last 30 years. Recently, a project to automate the construction of these databases was launched. The automated processing chain was developed in R and includes four main steps: - the automatic delineation of catchments boundaries from a digital elevation model by coupling R with GRASS GIS (rgrass7 package) and FORTRAN codes - the estimation of climate inputs at the catchment-scale by crossing catchment boundaries and climatic grids (raster package) - the use of a "shiny" interface including dynamic leaflet maps and real time upstream drainage basins to ease the expertise on the exact location of hydrometric stations - the production of synthetic sheets cross-referencing various hydrological, topographical and climatic data at the basin scale with dynamic graphs (dygraphs package) and static graphics (available at http://webgr.irstea.fr/activites/base-de-donnees/) This presentation will show how the developments made in R and existing packages were combined and implemented for building a large database at the national scale for research and operational applications in hydrology.