Single-walled carbon nanotubes (SWNTs) possess unique electronic properties that make them very promising materials for use in both nano-electronics and thin film devices. However, SWNTs are always produced as a mixture of metallic and semiconducting nanotubes, which is a major roadblock to their widespread application. This tutorial review provides a brief summary of ways of separating single-walled carbon nanotubes into metallic and semiconducting fractions. Various methods including selective growth, selective removal, selective adsorption and band structure modulation--all of which aim to produce pure SWNTs with well-defined electronic properties--are systematically discussed. The main problems in this field, the outlook for separation techniques and some views of future developments are presented.