US 8,527,478 B1Granted

Handling bulk and incremental updates while maintaining consistency

AssigneeGoogle Inc.

InventorsSeth Goldman, David Eustis, Tudor Hulubei, John Danaher

Life of this patent

Estimating the term…

Loading patent…

Abstract

Aspects of the invention relate generally to updating of interconnected features stored in a repository while maintaining highly consisting views of the repository data. More specifically, the invention relates to updating some features in bulk while providing for incremental updates for other features which are not completely independent of the bulk edits. The bulk edits are grouped into stages or “epochs” associated with various manageable states, including, for example, a building state, an approval state, a live state, and an abandoned state. The current view of the data is thus based on the status of each of the epochs. At the same time, incremental edits may be performed where the incremental edits are independent of the current live bulk epoch. This allows the system to maintain consistent views of the data until an entire group of features moves into the live state.

Claims18

Dependent claims

A computer-implemented method of updating repository data, the method comprising:

storing a plurality of first features in a first partition of a data repository, each first feature including a feature identifier that identifies that respective first feature;
storing a plurality of second features in a second partition of the repository, each second feature including a feature identifier that identifies that respective second feature and a subset of the plurality of second features including references identifying a feature identifier of a particular first feature of the plurality of first features;
generating a table of edit data including edits to the plurality of first features, each edit including an associated feature identifier of the respective first features, the edits to the plurality of first features being associated with a value and the table including a status column describing a status of the edits for a particular value as in an editing status where the edits to the plurality of first features are not accessible by client devices;
generating a second version of each of the second features by updating the references identifying the feature identifier of the particular first feature based on the value and the table of edits;
storing the second versions of each of the second features in the second partition of the repository;
and
updating the status column to a live state where the edits for the particular value are accessible by client devices.

4 more elements

A computer comprising:

memory including a first partition for storing a plurality of first features, each first feature including a feature identifier that identifies that respective first feature, the memory also including a second partition for storing a plurality of second features, each second feature including a feature identifier that identifies that respective second feature and a subset of the plurality of second features including references identifying a feature identifier of a particular first feature of the plurality of first features;
and
a processor operable to:
generate a table of edit data including edits to the plurality of first features, each edit including an associated feature identifier of the respective first features, the edits to the plurality of first features being associated with a value and the table including a status column describing a status of the edits for a particular value as in an editing status where the edits to the plurality of first features are not accessible by client devices;
generate a second version of each of the second features by updating the references identifying the feature identifier of the particular first feature based on the value and the table of edits;
store the second versions of each of the second features in the second partition of the repository;
and
update the status column to a live state where the edits for the particular value are accessible by client devices.

5 more elements

Dependencies are inferred from the claim text.