公开(公告)号：US11620280B2

公开(公告)日：2023-04-04

申请号：US17444715

申请日：2021-08-09

Applicant: Palantir Technologies Inc.

Inventor： Benjamin Duffield ,  Joshua Casale ,  Mark Elliot ,  Matthew Sills ,  Robert Kruszewski ,  Rahij Ramsharan

IPC: G06F16/23 ,  G06F16/27 ,  G06F16/2455 ,  G06F16/248

Abstract: A database system comprised of a decoupled compute layer and storage layer is implemented to store, build, and maintain a canonical dataset, a temporary buffer, and projection datasets. The canonical dataset is a set of batch updated data. The data is appended in chunks to the canonical dataset such that the canonical dataset becomes a historical dataset over time. The buffer is a write ahead log that contains the most recent chunks of data and provides atomicity and durability for the database system. The projection datasets are indexes of the canonical dataset and/or the buffer that may have single or multiple column sort-orders and/or particular data formats. The writes to the canonical dataset, projection datasets, and buffer may be asynchronous and therefore the database system is advantageously less resource constrained.

公开(公告)号：US20210109983A1

公开(公告)日：2021-04-15

申请号：US17066922

申请日：2020-10-09

Applicant: Palantir Technologies Inc.

Inventor： Elliot Colquhoun ,  Hugh Zabriskie ,  Joshua Casale ,  Rahij Ramsharan

IPC: G06F16/9532 ,  G06F16/9538 ,  G06F16/951 ,  G06F16/954

Abstract: A method, performed by one or more processors, including: receiving a user query; deriving, based on the user query, a search engine query for use with search engine software and a data access query for use with data access software; sending, to the search engine software, the search engine query to cause the search engine software to locate first data stored in a search engine index; sending, to the data access software, the data access query to cause the data access software to locate second data stored in a distributed data store; receiving, from the search engine software, the first data; and receiving, from the data access software, the second data.

公开(公告)号：US11928166B2

公开(公告)日：2024-03-12

申请号：US17901196

申请日：2022-09-01

Applicant: Palantir Technologies Inc.

Inventor： Elliot Colquhoun ,  Hugh Zabriskie ,  Joshua Casale ,  Rahij Ramsharan ,  Rushad Heerjee

IPC: G06F7/02 ,  G06F16/00 ,  G06F16/951 ,  G06F16/9532 ,  G06F16/9538 ,  G06F16/954 ,  G06F16/13

CPC classification number: G06F16/9532 ,  G06F16/951 ,  G06F16/9538 ,  G06F16/954 ,  G06F16/134

Abstract: A method, performed by one or more processors, including: receiving a user query; deriving, based on the user query, a search engine query for use with search engine software and a data access query for use with data access software; sending, to the search engine software, the search engine query to cause the search engine software to locate first data stored in a search engine index; sending, to the data access software, the data access query to cause the data access software to locate second data stored in a distributed data store; receiving, from the search engine software, the first data; and receiving, from the data access software, the second data.

公开(公告)号：US20240248939A1

公开(公告)日：2024-07-25

申请号：US18435740

申请日：2024-02-07

Applicant: Palantir Technologies Inc.

Inventor： Elliot Colquhoun ,  Hugh Zabriskie ,  Joshua Casale ,  Rahij Ramsharan ,  Rushad Heerjee

IPC: G06F16/9532 ,  G06F16/13 ,  G06F16/951 ,  G06F16/9538 ,  G06F16/954

CPC classification number: G06F16/9532 ,  G06F16/951 ,  G06F16/9538 ,  G06F16/954 ,  G06F16/134

Abstract: A method, performed by one or more processors, includes: receiving, from search engine software, first data. In some examples, the first data includes one or more data records. In some examples, the method further includes deriving, based on the first data, a data access query for locating second data using data extracted from the one or more data records. In some examples, the second data includes one or more data objects. In some examples, the method further includes sending, to data access software, the data access query to cause the data access software to locate the second data stored in a second data store different from a first data store storing the first data.

公开(公告)号：US20230004606A1

公开(公告)日：2023-01-05

申请号：US17901196

申请日：2022-09-01

Applicant: Palantir Technologies Inc.

Inventor： Elliot Colquhoun ,  Hugh Zabriskie ,  Joshua Casale ,  Rahij Ramsharan ,  Rushad Heerjee

IPC: G06F16/9532 ,  G06F16/951 ,  G06F16/9538 ,  G06F16/954

Abstract: A method, performed by one or more processors, including: receiving a user query; deriving, based on the user query, a search engine query for use with search engine software and a data access query for use with data access software; sending, to the search engine software, the search engine query to cause the search engine software to locate first data stored in a search engine index; sending, to the data access software, the data access query to cause the data access software to locate second data stored in a distributed data store; receiving, from the search engine software, the first data; and receiving, from the data access software, the second data.

公开(公告)号：US20230281191A1

公开(公告)日：2023-09-07

申请号：US18178294

申请日：2023-03-03

Applicant: Palantir Technologies Inc.

Inventor： Benjamin Duffield ,  Joshua Casale ,  Mark Elliot ,  Matthew Sills ,  Robert Kruszewski ,  Rahij Ramsharan

IPC: G06F16/23 ,  G06F16/27 ,  G06F16/2455 ,  G06F16/248

CPC classification number: G06F16/2379 ,  G06F16/273 ,  G06F16/2455 ,  G06F16/248

Abstract: A database system comprised of a decoupled compute layer and storage layer is implemented to store, build, and maintain a canonical dataset, a temporary buffer, and projection datasets. The canonical dataset is a set of batch updated data. The data is appended in chunks to the canonical dataset such that the canonical dataset becomes a historical dataset over time. The buffer is a write ahead log that contains the most recent chunks of data and provides atomicity and durability for the database system. The projection datasets are indexes of the canonical dataset and/or the buffer that may have single or multiple column sort-orders and/or particular data formats. The writes to the canonical dataset, projection datasets, and buffer may be asynchronous and therefore the database system is advantageously less resource constrained.

公开(公告)号：US11468130B2

公开(公告)日：2022-10-11

申请号：US17066922

申请日：2020-10-09

Applicant: Palantir Technologies Inc.

Inventor： Elliot Colquhoun ,  Hugh Zabriskie ,  Joshua Casale ,  Rahij Ramsharan ,  Rushad Heerjee

IPC: G06F7/02 ,  G06F16/00 ,  G06F16/9532 ,  G06F16/951 ,  G06F16/9538 ,  G06F16/954 ,  G06F16/13

Abstract: A method, performed by one or more processors, including: receiving a user query; deriving, based on the user query, a search engine query for use with search engine software and a data access query for use with data access software; sending, to the search engine software, the search engine query to cause the search engine software to locate first data stored in a search engine index; sending, to the data access software, the data access query to cause the data access software to locate second data stored in a distributed data store; receiving, from the search engine software, the first data; and receiving, from the data access software, the second data.

公开(公告)号：US20220058183A1

公开(公告)日：2022-02-24

申请号：US17444715

申请日：2021-08-09

Applicant: Palantir Technologies Inc.

Inventor： Benjamin Duffield ,  Joshua Casale ,  Mark Elliot ,  Matthew Sills ,  Robert Kruszewski ,  Rahij Ramsharan

IPC: G06F16/23 ,  G06F16/27 ,  G06F16/2455 ,  G06F16/248

Abstract: A database system comprised of a decoupled compute layer and storage layer is implemented to store, build, and maintain a canonical dataset, a temporary buffer, and projection datasets. The canonical dataset is a set of batch updated data. The data is appended in chunks to the canonical dataset such that the canonical dataset becomes a historical dataset over time. The buffer is a write ahead log that contains the most recent chunks of data and provides atomicity and durability for the database system. The projection datasets are indexes of the canonical dataset and/or the buffer that may have single or multiple column sort-orders and/or particular data formats. The writes to the canonical dataset, projection datasets, and buffer may be asynchronous and therefore the database system is advantageously less resource constrained.