This survey provides a comprehensive review of knowledge graphs (KGs), covering representation learning, acquisition, completion, temporal KGs, and knowledge-aware applications. It summarizes recent breakthroughs and outlines future research directions. The paper proposes a full-view categorization and new taxonomies for these topics. Knowledge graph embedding is organized from four aspects: representation space, scoring function, encoding models, and auxiliary information. For knowledge acquisition, especially knowledge graph completion, embedding methods, path inference, and logical rule reasoning are reviewed. Emerging topics such as meta relational learning, commonsense reasoning, and temporal KGs are also explored. The paper provides a curated collection of datasets and open-source libraries for different tasks. It concludes with a thorough outlook on promising research directions. Knowledge graphs represent structured facts with entities, relationships, and semantic descriptions. They are used in AI for reasoning and problem-solving. Recent advances focus on knowledge representation learning (KRL) or knowledge graph embedding (KGE) by mapping entities and relations into low-dimensional vectors. Tasks include knowledge graph completion (KGC), triple classification, entity recognition, and relation extraction. Knowledge-aware models benefit from heterogeneous information, ontologies, and multi-lingual knowledge, enabling applications like recommendation systems and question answering. Examples include Microsoft's Satori and Google's Knowledge Graph. The survey reviews KRL from four scopes: representation space, scoring function, encoding models, and auxiliary information. It categorizes knowledge acquisition into KGC, relation extraction, and entity discovery. Temporal KGs incorporate temporal information. Knowledge-aware applications include NLU, QA, and real-world tasks. The paper also discusses related surveys and provides a detailed overview of KRL, including representation spaces like point-wise, complex vector, Gaussian distribution, and manifold space. Scoring functions include distance-based and similarity-based methods. Encoding models include linear/bilinear models, factorization models, and neural networks. Auxiliary information such as textual descriptions, type information, visual information, and uncertain information are also considered. The survey concludes with a summary of recent KRL models and future research directions.This survey provides a comprehensive review of knowledge graphs (KGs), covering representation learning, acquisition, completion, temporal KGs, and knowledge-aware applications. It summarizes recent breakthroughs and outlines future research directions. The paper proposes a full-view categorization and new taxonomies for these topics. Knowledge graph embedding is organized from four aspects: representation space, scoring function, encoding models, and auxiliary information. For knowledge acquisition, especially knowledge graph completion, embedding methods, path inference, and logical rule reasoning are reviewed. Emerging topics such as meta relational learning, commonsense reasoning, and temporal KGs are also explored. The paper provides a curated collection of datasets and open-source libraries for different tasks. It concludes with a thorough outlook on promising research directions. Knowledge graphs represent structured facts with entities, relationships, and semantic descriptions. They are used in AI for reasoning and problem-solving. Recent advances focus on knowledge representation learning (KRL) or knowledge graph embedding (KGE) by mapping entities and relations into low-dimensional vectors. Tasks include knowledge graph completion (KGC), triple classification, entity recognition, and relation extraction. Knowledge-aware models benefit from heterogeneous information, ontologies, and multi-lingual knowledge, enabling applications like recommendation systems and question answering. Examples include Microsoft's Satori and Google's Knowledge Graph. The survey reviews KRL from four scopes: representation space, scoring function, encoding models, and auxiliary information. It categorizes knowledge acquisition into KGC, relation extraction, and entity discovery. Temporal KGs incorporate temporal information. Knowledge-aware applications include NLU, QA, and real-world tasks. The paper also discusses related surveys and provides a detailed overview of KRL, including representation spaces like point-wise, complex vector, Gaussian distribution, and manifold space. Scoring functions include distance-based and similarity-based methods. Encoding models include linear/bilinear models, factorization models, and neural networks. Auxiliary information such as textual descriptions, type information, visual information, and uncertain information are also considered. The survey concludes with a summary of recent KRL models and future research directions.