In order to standardize the information releasing behaviors， electricity regulatory authorities require the power dispatching center and the electricity trade center to release lots of market-related information. From the aspect of data learning， huge mount of data might not provide valuable information of electricity trade. In contrast， it possibly causes the data disaster. This paper quantitatively and qualitatively studies the category method and releasing criteria of electricity market information. Based on the optimal criteria of market-based electricity trade， the concept of key information set and the fairness criteria of electricity trade are proposed. The key information set is the minimum data set required to check the fairness of electricity trade. Using the key information set， redundant data can be eliminated from released information to avoid the data disaster. Numeric studies also show that market clearing price is less valuable than nodal marginal price， which contains lots of electricity trade information and should be one of the most important parts of released information in electricity market.