Due to the mass availability of textual data on Web, text classification (TC), classifying texts into predetermined sets becomes a spotlight for researchers. A number of TC applications have been proposed yet very few studies reported an overview of TC research area in a proper and systematic manner. This paper aims to provide an overview of TC research trends and gaps by structuring and analyzing research patterns, encountered problems and problem-solving methods in TC. In other words, this study highlights problem types, data sources, choice of language of text and types of applied techniques in TC. An intensive systematic study is conducted by applying guidelines proposed by Petersen and colleagues in 2007. In this paper, ninety-six literatures from five electronic databases from 2006 to 2017 were systematically reviewed and followed each and every step properly in accordance with systematic mapping study. Nine main problems in TC research area were identified and significant findings which highlighted the evolution of TC research within the past 12 years were investigated. Different from other review articles, this paper highlighted issues and technical gaps of TC area in a useful and effective manner.