Data loading method for a data warehouse
AssigneeHewlett-Packard Development Company, L.P.
InventorsMadan Gopal Devadoss, Abhilash R, Panish Ramakrishna
Life of this patent
Estimating the term…
Abstract
Provided is a method of loading data in a data warehouse. Two types of work tables are generated, namely, a first work table and a second work table. The data is loaded into the first work table and prepared for loading, using the second work table as temporary table for data preparation. The prepared data is loaded into the data warehouse.
Claims10
A computer-implemented method of data loading in a data warehouse, comprising:
- generating two types of work tables comprising a first work table and a second work table, wherein the second work table is a temporary table for data preparation;
- loading data into the first work table;
- preparing the data in the first work table for loading into the data warehouse, wherein preparing the data includes:
- inserting data from the first work table into the second work table;
- updating legacy columns of the second work table using business key column values;
- deleting duplicates within the second work table using a business key column in the second work table;
- updating a surrogate key column of the second work table using business key columns in the data warehouse and the second work table;
- and
- updating foreign key reference column of the first work table using the updated surrogate key of the second work table;
- and
- loading the prepared data in the data warehouse.
A system, comprising:
- a processor and a memory;
- a data loader module, implemented by the processor, which:
- generates two types of work tables comprising a first work table and a second work table, wherein the second work table is a temporary table for data preparation;
- loads data into the first work table;
- prepares the data in the first work table for loading into a data warehouse stored in the memory, wherein to prepare the data, the processor performs:
- inserting data from the first work table into the second work table;
- updating legacy columns of the second work table using business key column values;
- deleting duplicates within the second work table using a business key column in the second work table;
- updating a surrogate key column of the second work table using business key columns in the data warehouse and the second work table;
- and
- updating foreign key reference column of the first work table using the updated surrogate key of the second work table;
- and
- loads the prepared data in the data warehouse.
A non-transitory machine-readable medium comprising instructions to implement a method of loading data into a data warehouse, wherein the instructions, when executed by a processor, cause the processor to perform:
- generating two types of work tables comprising a first work table and a second work table, wherein the second work table is a temporary table for data preparation;
- loading data into the first work table;
- preparing the data in the first work table for loading into a data warehouse stored in the memory, wherein to prepare the data, the processor performs:
- inserting data from the first work table into the second work table;
- updating legacy columns of the second work table using business key column values;
- deleting duplicates within the second work table using a business key column in the second work table;
- updating a surrogate key column of the second work table using business key columns in the data warehouse and the second work table;
- and
- updating foreign key reference column of the first work table using the updated surrogate key of the second work table;
- and
- loading the prepared data in the data warehouse.
Dependencies are inferred from the claim text.