Online hate speech (OHS) on social networking sites such as Facebook, Instagram, Twitter(now X), WhatsApp, Reddit, etc. has emerged as a growing concern due to its negative psychological and emotional effects on the people and society. This issue is amplified by the enhancement of voices, misinformation, anonymity, and the lack of accountability on social media, where users may not face immediate or tangible consequences for their words. Recognising the detrimental impact of OHS on individual well-being, there is growing interest in developing effective mechanisms for its detection and mitigation. This study presents a comprehensive survey of existing literature on OHS detection for multimodal and multilingual data available on social networking sites. The analysis involves examining available datasets and reviewing 23 most recent research papers published between year 2020 and 2024. The surveyed papers are categorised based on the datasets and approaches used such as traditional machine learning, deep learning algorithms, and other approaches like rule-based and hybrid methods are also explored.