The internet of things (IoT) is the networking of interrelated devices and sensors connected through the internet to transfer and share data. The data gathered from these devices may have anomalies or other errors for various reasons, such as malicious activities or sensor failures. Anomaly detection is found in several domains, such as fault detection and health monitoring systems. In this paper, we review and analyze the relevant literature on existing anomaly detection techniques that apply different machine learning approaches in the IoT. In addition, we examine different anomaly detection datasets used in IoT and highlight the most concerning issues with these datasets for different approaches, and list several future research directions. We believe this survey will serve as a starting point for researchers to gain knowledge from the IoT that employs machine learning approaches to detect anomalies. Moreover, the datasets that associate with anomaly detection in IoT, issues and future directions from a dataset perspective.