The invention relates to the technical field of intelligent
data query, in particular to a
natural language intelligent number asking method, device and equipment and a computer storage medium. According to the
natural language intelligent number asking method, firstly, a local pre-training
language model subjected to field corpus fine adjustment is used for achieving accurate matching from query elements to
library table indexes, and then
SQL query meta-languages are generated based on the
library table indexes obtained through matching and original query texts. The method comprises the steps that firstly,
SQL query meta-languages are obtained, column sorting is conducted according to
data query results returned by the
SQL query meta-languages,
data analysis texts are generated, finally, correlation indexes are calculated according to
library table indexes in the current SQL query meta-languages, the
data query results are output, and the
data analysis texts and the correlation indexes serve as final query results; according to the method, the industry index elements in the
natural language query statement are accurately recognized, the
structured analysis capability of the query elements is improved, and finally the query accuracy is improved.