Method and device for establishing NoSQL database index for semi-structured data

The invention provides a method for establishing an NoSQL database index for semi-structured data. The method comprises the steps of preprocessing semi-structured source data; storing text segments obtained after preprocessing in a data sheet, wherein the data sheet comprises a first major key assembly including a structured thread major key and a sequential value major key, a structured thread is divided into multiple continuous intervals according to a determined sequence, and a specific key value is allocated to each interval to serve as the key values of the structured thread major key; establishing a reverse index table for the text segments obtained after preprocessing, wherein the reverse index table comprises a second major key assembly including a structured thread major key and a keyword major key, and relevant text segment sequence tags are recorded to serve as index values corresponding to the key values of the major keys. The index values with the same keyword major key and different structured thread major key values are located on different lines in the reverse index table. In this way, database index query efficiency is improved, and updating is facilitated. Furthermore, the invention provides a device for establishing an NoSQL database index for semi-structured data.