US 12,111,806 B2
Method and apparatus for determining associated data linkages
Mridul Jain, Cupertino, CA (US); Saigopal Thota, Fremont, CA (US); Xun Luan, Santa Clara, CA (US); and Gajendra Alias Nishad Kamat, Cupertino, CA (US)
Assigned to Walmart Apollo, LLC, Bentonville, AR (US)
Filed by Walmart Apollo, LLC, Bentonville, AR (US)
Filed on Jan. 31, 2019, as Appl. No. 16/263,545.
Prior Publication US 2020/0250161 A1, Aug. 6, 2020
Int. Cl. G06F 16/22 (2019.01); G06F 9/455 (2018.01); G06F 16/21 (2019.01)
CPC G06F 16/22 (2019.01) [G06F 9/45558 (2013.01); G06F 16/211 (2019.01); G06F 2009/45595 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising: a computing device comprising at least one processor and configured to:
obtain linkage data identifying a plurality of links for a plurality of nodes, wherein each link directly associates at least two of the plurality of nodes: a child node and a parent node having a hierarchy immediately above a hierarchy of the child node;
partition the linkage data into a first number of logical partitions;
execute, for each of the first number of logical partitions, a union find algorithm wherein the execution of each union find algorithm assigns an identifier to each node of each link based on the linkage data of each respective logical partition;
execute a global shuffle algorithm that reassigns at least one node,
wherein each of the at least one node has multiple node copies that were linked to different parent nodes respectively in multiple different logical partitions after executing the union find algorithm and before executing the global shuffle algorithm,
wherein each of the at least one node is reassigned based on:
reassigning the multiple node copies from the multiple different logical partitions to a same logical partition of the multiple different logical partitions,
re-linking the multiple node copies, which were linked to different parent nodes in the multiple different logical partitions before executing the global shuffle algorithm, to a same parent node in the same logical partition by executing the global shuffle algorithm; and
execute a path compression algorithm that compresses at least one multiple-hop link, each of which associates multiple nodes in different hierarchies respectively with multiple hops, wherein each of the at least one multiple-hop link is compressed based on:
identifying a final parent node in a highest hierarchy among the multiple nodes,
re-linking the remaining nodes other than the final parent node in the multiple-hop link to the final parent node directly to form multiple single-hop links.