This dataset contains anonymized doctor-patient consultation records collected from a widely used online telemedicine platform in China. It is specifically curated for natural language processing (NLP) applications in the healthcare domain, particularly focusing on diabetes-related medical dialogues. The dataset has been carefully segmented, translated, and organized into two main components: 1. original_txt.zip Contains 19 individual .txt files named part_1.txt through part_19.txt. Each file includes 1,000 real-world consultation threads in Chinese, with structured multi-turn dialogues between patients and doctors. All entries follow a consistent format, segmented by markers such as id=xxx, to facilitate automated parsing and processing. Personally identifiable information (PII) has been strictly removed to ensure privacy. 2. translated_txt.zip Contains the English-translated versions of each corresponding file in original_txt.zip, following a one-to-one mapping (part_1_english.txt, part_2_english.txt, etc.). Translation was performed using Google Translate (via googletrans library), preserving the original structure, labeling, and dialogue format for each entry. These English versions enable broader accessibility and cross-linguistic research.