Enabling users to query databases with natural language has been an active research area for many decades. This chapter presents the problem of translating natural language questions to SQL and the challenges it poses. It presents an overview of early efforts that did not rely on neural networks, their architecture, and the main techniques that they relied on. Finally, it focuses on setting the scene for the era of neural Text-to-SQL systems that has gathered significant research attention. To achieve this, an in-depth analysis of available datasets is presented, highlighting the main categories of datasets and analyzing the most important ones. Additionally, the metrics that are used to evaluate neural Text-to-SQL systems are presented, explaining the importance and the weaknesses of modern evaluation strategies.