1. Collect data and clean up the collected data to ensure the accuracy and consistency of the data. Taking market data as an example, accuracy means that the collected data need to be repeatedly compared and confirmed, and the collected data are regularly compared and tested to ensure the accuracy of the data needed for strategy formation. Consistency means that the data of strategy test and strategy realization need to be homologous, and the market data is sent from the exchange to all futures companies. Due to the local time setting and receiving delay of the receiving server of futures companies, the market data forwarded by different futures companies may be more or less inconsistent, so the consistency of data needs to be considered in the process of database construction. The data used for strategy history testing and statistics should be consistent with the data used for quantitative trading strategy implementation in the future, so that quantitative strategy testing can be more reliable.

2. The realized data need the same source, and the market data is sent from the exchange to the futures company. Due to the local time setting and receiving delay of the receiving server of futures companies, the market data forwarded by different futures companies may be more or less inconsistent, so the consistency of data needs to be considered in the process of database construction. The data used for strategy history testing and statistics should be consistent with the data used for quantitative trading strategy implementation in the future, so that quantitative strategy testing will be more reliable.

3. The database needs to integrate some algorithmic functions. These functions include simple general functions, numerical analysis, statistics, data access and related functions of financial securities. Some high-end database construction integrates artificial intelligence algorithms, such as genetic algorithm, ant colony algorithm and support vector machine. This part is relatively high-end and has high requirements for general financial institutions. Direct integration into the database is a big algorithm project, and some institutions use the method of calling external algorithm engines and returning results to realize artificial intelligence algorithms. However, with more and more talents in mathematics, physics and computer entering the research of quantitative strategy, many institutions have integrated such algorithms in their own databases, which has promoted the simple database to the research and development platform of quantitative strategy.