In the ubiquitous electric power internet of things（UEP-IoT）

the traditional cloud computing operation mode that uses centralized processing mode in the power data center is difficult to deal with the rapidly expanding data scale and requires high computational capacity for the power data center. Aiming at solving the above problems

in this paper

based on designing an edge computing framework for processing the UEP-Io T data

a novel adaptive data processing method is proposed. Firstly

via design a counting bloom filter-based adaptively sensing mechanism for the edge nodes

the UEP-Io T data categories and the corresponding data segments can be identified automatically. Secondly

an edge-node-aware adaptive data processing scheme

containing a universal portable end-of-file（EOF） identification plug-in and a data-copy management system

is designed to realize the UEP-IoT data collection. The numerical examples show that

for the same data set

the accuracy of the data type identification of our proposed method can be improved by over1.3%. Besides

compared with the traditional data replica management scheme

the average processing time of the proposed combining local management with cluster management scheme for the data set decreases by 33.33%

respectively

which shows that the proposed scheme can effectively improve the processing performance of massive UEP-IoT data.