This paper provides a comprehensive survey of person re-identification (Re-ID) systems, focusing on both closed-world and open-world settings. The authors categorize the existing methods into three main components: feature representation learning, deep metric learning, and ranking optimization. They provide an in-depth analysis of these components, discussing their advantages and limitations. The paper also introduces a new evaluation metric, mINP, which measures the cost of finding all correct matches, and a powerful baseline, AGW, that achieves state-of-the-art performance on twelve datasets for four different Re-ID tasks. The authors highlight the challenges in practical applications, such as heterogeneous data, raw images/videos, limited or noisy annotations, and open-set scenarios, and discuss several under-investigated open issues. The survey covers a wide range of datasets and state-of-the-art methods, providing insights into the current state of the field and suggesting future research directions.This paper provides a comprehensive survey of person re-identification (Re-ID) systems, focusing on both closed-world and open-world settings. The authors categorize the existing methods into three main components: feature representation learning, deep metric learning, and ranking optimization. They provide an in-depth analysis of these components, discussing their advantages and limitations. The paper also introduces a new evaluation metric, mINP, which measures the cost of finding all correct matches, and a powerful baseline, AGW, that achieves state-of-the-art performance on twelve datasets for four different Re-ID tasks. The authors highlight the challenges in practical applications, such as heterogeneous data, raw images/videos, limited or noisy annotations, and open-set scenarios, and discuss several under-investigated open issues. The survey covers a wide range of datasets and state-of-the-art methods, providing insights into the current state of the field and suggesting future research directions.