Systems and methods for generating a data definition language statement
AssigneeRocket Software, Inc.
InventorsTom Ulveman Jensen, Revanth Akella, Lester Manalastas, Rekha Gadikota
Life of this patent
Estimating the term…
Abstract
Systems and methods for generating a data definition language (DDL) statement for a database management product. The method includes receiving at an interface a partial DDL text input; providing the partial DDL text input to a machine learning model trained on a corpus of DDL statements and configured to analyze the received partial DDL text input and generate a complete DDL statement based on the analysis of the received partial DDL text input and the corpus of DDL statements.
Claims20
A method for generating a data definition language (DDL) statement for a database management product, the method comprising:
- receiving at an interface a partial DDL text input;
- providing the partial DDL text input to a machine learning model trained on a corpus of DDL statements and trained by:
- receiving a plurality of sample DDL statements,
- identifying an object type associated with each of the sample DDL statements,
- creating sequences of tokens representing the objects, wherein the length of the sequences is based on the identified object type, and
- providing the sequences of tokens to the machine learning model for training,
- wherein the trained machine learning model then
- analyzes the received partial DDL text input, and
- generates a complete DDL statement based on the analysis of the received partial DDL text input and the corpus of DDL statements.
A system for generating a data definition language (DDL) statement for a database management product, the system comprising:
- an interface for at least receiving a partial DDL text input from a user;
- a database for storing a corpus of DDL statements;
- and
- one or more processors executing instructions stored on memory and configured to:
- execute a machine learning model to analyze the corpus of DDL statements,
- wherein the machine learning model is trained by:
- receiving a plurality of sample DDL statements,
- identifying an object type associated with each of the sample DDL statements,
- creating sequences of tokens representing the objects, wherein the length of the sequences is based on the identified object type and the machine learning model is trained on the sequences of tokens, and
- generate a complete DDL statement based on the analysis of the received partial DDL text input and the corpus of DDL statements, wherein the interface is further configured to present the generated complete DDL statement to the user.
The system of 11 wherein the one or more processors generate the complete DDL statement by predicting at least one text item to follow the partial DDL text input.
A computer program product for generating a data definition language (DDL) statement for a database management product, the computer program product comprising computer executable code embodied in one or more non-transitory computer readable media that, when executing on one or more processors, performs the steps of:
- receiving at an interface a partial DDL text input;
- providing the partial DDL text input to a machine learning model trained on a corpus of DDL statements and trained by:
- receiving a plurality of sample DDL statements,
- identifying an object type associated with each of the sample DDL statements,
- creating sequences of tokens representing the objects, wherein the length of the sequences is based on the identified object type, and
- providing the sequences of tokens to the machine learning model for training, wherein the trained machine learning model then:
- analyzes the received partial DDL text input, and
- generates a complete DDL statement based on the analysis of the received partial DDL text input and the corpus of DDL statements.
Dependencies are inferred from the claim text.