Abstract: The measurement data of distribution transformers has high business value, but due to multiple factors, there are a few flaws in their integrity. Traditional data completion techniques such as state estimation and interpolation ignore the data loss caused by power outage in distribution transformers, which will affect the accuracy of business analysis. On the basis of multi-source data fusion in operation, distribution, and dispatch, this paper proposes a framework and method for identifying integrity anomalies in distribution transformer data. Firstly, breakpoint fragments are generated based on the missing fragments of distribution transformer data, and preliminary analysis is conducted after filtering for jitter anomalies, based on the trend of electricity growth rate on the similar day. Secondly, breakpoint fragments are aggregated into breakpoint events according to the spatiotemporal correlation characteristics. Finally, based on the core business perspectives of distribution network such as power outage and line loss, the power outage event is verified and the missing distribution transformer data caused by power outage is annotated. Practical cases have proven the effectiveness of the above methods. This technology has been applied in engineering in Jiangsu, significantly improving the quality of distribution transformer data. Furthermore, a good interactive model of promoting governance through use has been formed, promoting the deepening application of existing business systems such as line loss and reliability management.