Lakes are considered as “islands” for assessing microbial biogeography, but bacterial community composition (BCC) and function may vary significantly within lakes, with the roles of scale and connectivity still unclear. This study investigated the spatial and temporal heterogeneity of the BCC (Automated Ribosomal Intergenic Spacer Analysis) and Functional Traits (FT, carbon-source utilization), and the contribution of: (i) environmental variables, (ii) phytoplankton, (iii) season and (iv) space, through variance partitioning in the large and well-mixed Lake Mangueira. The BCC and FT differed in time and space, with BCC being explained by environmental variables and phytoplankton, whereas FT explained only by space. The smallest scale of variability detected for the BCC and FT (~49 km) was the largest in comparison with findings in other studies, suggesting an effect of lake size (fetch and connectivity). Our results indicate that barriers to bacterial dispersal due to long distances are overcome by high connectivity, reinforcing the role of species sorting for BCC. FT were probably driven by gene dispersal and/or the effects of local conditions on migrant bacterial taxa and resuspended bacteria. Our results highlight the role of within-lake heterogeneity for ecosystem functioning and the implications for the appropriate scale for sampling bacterial communities.