# US 2026/0099476 A1 · Reading both pages and segments via applying a columnar-formatted structuring-based read process

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0099476 A1
- Document ID: US20260099476A1
- Status: Published application
- Kind: A1
- Application number: 18/910,353
- Filed: 2024-10-09
- Published: 2026-04-09
- Earliest claimed: 2024-10-09
- Assignees: Ocient Holdings LLC
- Applicants: Ocient Holdings LLC
- Inventors: Greg R. Dhuse; Benjamin Daniel Rabe; Finley Jordan Lau; Pieter Charles Jas Svenson
- Web page: https://patentgrep.com/patent/US20260099476A1

## CPC

- G06F 16/221: {Column-oriented storage; Management thereof}
- G06F 16/24532: {of parallel queries}
- G06F 16/2455: Query execution
- G06F 16/258: {Data format conversion from or to a database}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A database system is operable to generating a plurality of pages from a plurality of rows, where each page of the plurality of pages is generated to include a corresponding set of rows of the plurality of rows in accordance with a columnar-formatted structuring based on generating a plurality of columns from the corresponding set of rows. A query is executed based on reading a first set of column values for a first plurality of rows based on applying a columnar-formatted structuring-based read process and reading a second set of column values for a second plurality of rows based on applying the columnar-formatted structuring-based read process. At least one page conversion process is performed to generate an additional plurality of segments based on reading all column values for the plurality of rows based on applying the columnar-formatted structuring-based read process.

## Claims

### Claim at position 1

1-20. (canceled)

### Claim 21

A parallelized data input sub-system of a database system, the parallelized data input sub-system comprises:
a plurality of computing device clusters, wherein a computing device cluster of the plurality of computing device clusters includes a plurality of computing devices, wherein the plurality of computing device clusters includes pluralities of computing nodes, wherein a plurality of the pluralities of computing nodes is configured as loader computing nodes, wherein the loader computing nodes are operable, in parallel, to receive respective portions of a dataset, wherein a loader computing node of the loader computing nodes is operable to:
receive rows of columnar data of a first portion of the portions of the dataset;
convert a page of the rows of columnar data into a page of column oriented data;
temporarily store the page of column oriented data;
convert the page of column oriented data into long term storage data, wherein the long term storage data has a corresponding column orientation to the page of column oriented data; and
while receiving the first portion:
receive a query request regarding a specific page of column oriented data prior to the specific page of column oriented data being converted into specific long term storage data; and
when the specific page of columnar oriented data has been temporarily stored:
provide the specific page of column oriented data to a computing device of the database system executing the query on the specific page of column oriented data.

### Claim 22 (depends on 21)

The parallelized data input sub-system of claim 21, wherein the loader computing node is operable to convert the page of the rows of columnar data into the page of column oriented data by:
determining a key column of the page of rows of columnar data;
grouping the page of rows of columnar data by the key column to produce an ordered page of rows of columnar data;
rotating the ordered page of rows of columnar data to produce ordered columns of data; and
appending respective column key mappings and row identification mappings to respective columns of the ordered columns of data to produce the page of column oriented data.

### Claim 23 (depends on 22)

The parallelized data input sub-system of claim 22, wherein the loader computing node is further operable to:
compress the ordered columns of data to produce compressed ordered columns of data; and
appending the respective column key mappings, respective compression frame mappings, and the respective row identification mappings to the respective columns of the compressed ordered columns of data to produce the page of column oriented data.

### Claim 24 (depends on 21)

The parallelized data input sub-system of claim 21, wherein the loader computing node is further operable to:
receive second rows of columnar data of a second portion of the portions of the dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the second portion:
receive the query request regarding the specific page of column oriented data prior to the specific page of column oriented data being converted into the specific long term storage data; and
when the specific page of column oriented data has been temporarily stored:
provide the specific page of column oriented data to the computing device of the database system executing the query on the specific page of column oriented data.

### Claim 25 (depends on 21)

The parallelized data input sub-system of claim 21, wherein a second loader computing node of the loader computing nodes is operable to:
receive second rows of columnar data of a second portion of the portions of the dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the second portion:
receive the query request regarding the specific page of column oriented data prior to the specific page of column oriented data being converted into the specific long term storage data; and
when the specific page of column oriented data has been temporarily stored:
provide the specific page of column oriented data to the computing device of the database system executing the query on the specific page of column oriented data. second portion different node.

### Claim 26 (depends on 21)

The parallelized data input sub-system of claim 21 further comprises:
wherein the loader computing nodes are operable, in parallel, to receive respective second portions of a second dataset, wherein the loader computing node is operable to:
receive second rows of columnar data of a first portion of the second portions of the second dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the first portion of the second portions:
receive a second query request regarding a second specific page of column oriented data prior to the second specific page of column oriented data being converted into second specific long term storage data; and
when the second specific page of column oriented data has been temporarily stored:
provide the second specific page of column oriented data to a computing device of the database system executing the second query on the second specific page of column oriented data.

### Claim 27 (depends on 21)

The parallelized data input sub-system of claim 21, wherein the loader computing node is operable to convert the page of column oriented data into the long term storage data by:
redundancy encoding the page of column oriented data to produce a plurality of data and parity blocks;
generating one or more indexes regarding the page of column oriented data;
generating statistical data regarding the page of column oriented data; and
arranging the data and parity blocks, the one or more indexes, the statistical data, and a key column as the long term storage data.

### Claim 28 (depends on 27)

The parallelized data input sub-system of claim 27, wherein the loader computing node is further operable to:
include metadata regarding the data and parity blocks in the long term storage data.

### Claim 29

A computer readable memory comprises:
a first memory section that stores operational instructions that, when executed by loader computing nodes of a plurality of computing nodes of pluralities of computing nodes of a plurality of computing device clusters of a parallelized data input sub-system of a database system, cause the loader computing nodes to:
receive, in parallel, respective portions of a dataset;
a second memory section that stores operational instructions that, when executed by a loader computing node of the loader computing nodes, cause the loader computing node to:
receive rows of columnar data of a first portion of the portions of the dataset;
convert a page of the rows of columnar data into a page of column oriented data;
temporarily store the page of column oriented data;
convert the page of column oriented data into long term storage data, wherein the long term storage data has a corresponding column orientation to the page of column oriented data; and
while receiving the first portion:
receive a query request regarding a specific page of column oriented data prior to the specific page of column oriented data being converted into specific long term storage data; and
when the specific page of columnar oriented data has been temporarily stored:
provide the specific page of column oriented data to a computing device of the database system executing the query on the specific page of column oriented data.

### Claim 30 (depends on 29)

The computer readable memory of claim 29, wherein the second memory section further stores operational instructions that, when executed by the loader computing node, cause the loader computing node to convert the page of the rows of columnar data into the page of column oriented data by:
determining a key column of the page of rows of columnar data;
grouping the page of rows of columnar data by the key column to produce an ordered page of rows of columnar data;
rotating the ordered page of rows of columnar data to produce ordered columns of data; and
appending respective column key mappings and respective row identification mappings to respective columns of the ordered columns of data to produce the page of column oriented data.

### Claim 31 (depends on 30)

The computer readable memory of claim 30, wherein the second memory section further stores operational instructions that, when executed by the loader computing node, cause the loader computing node to:
compress the ordered columns of data to produce compressed ordered columns of data; and
appending the respective column key mappings, respective compression frame mappings, and the respective row identification mappings to the respective columns of the compressed ordered columns of data to produce the page of column oriented data.

### Claim 32 (depends on 29)

The computer readable memory of claim 29, wherein the second memory section further stores operational instructions that, when executed by the loader computing node, cause the loader computing node to:
receive second rows of columnar data of a second portion of the portions of the dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the second portion:
receive the query request regarding the specific page of column oriented data prior to the specific page of column oriented data being converted into the specific long term storage data; and
when the specific page of column oriented data has been temporarily stored:
provide the specific page of column oriented data to the computing device of the database system executing the query on the specific page of column oriented data.

### Claim 33 (depends on 29)

The computer readable memory of claim 29 further comprises:
a third memory section that stores operational instructions that, when executed by a second loader computing node of the loader computing nodes, cause the second loader computing node to:
receive second rows of columnar data of a second portion of the portions of the dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the second portion:
receive the query request regarding the specific page of column oriented data prior to the specific page of column oriented data being converted into the specific long term storage data; and
when the specific page of column oriented data has been temporarily stored:
provide the specific page of column oriented data to the computing device of the database system executing the query on the specific page of column oriented data. second portion different node.

### Claim 34 (depends on 29)

The computer readable memory of claim 29 further comprises:
wherein the first memory section further stores operational instructions that when executed by the loader computing nodes, cause the loader computing nodes to:
receive, in parallel, respective second portions of a second dataset; and
wherein the second memory section further stores operational instructions that when executed by the loader computing node, cause the loader computing node to:
receive second rows of columnar data of a first portion of the second portions of the second dataset;
convert a page of the second rows of columnar data into a second page of column oriented data;
temporarily store the second page of column oriented data;
convert the second page of column oriented data into second long term storage data, wherein the second long term storage data has the corresponding column orientation to the second page of column oriented data; and
while receiving the first portion of the second portions:
receive a second query request regarding a second specific page of column oriented data prior to the second specific page of column oriented data being converted into second specific long term storage data, and
when the second specific page of column oriented data has been temporarily stored:
provide the second specific page of column oriented data to a computing device of the database system executing the second query on the second specific page of column oriented data.

### Claim 35 (depends on 29)

The computer readable memory of claim 29, wherein the second memory section further stores operational instructions that, when executed by the loader computing node, cause the loader computing node to convert the page of column oriented data into the long term storage data by:
redundancy encoding the page of column oriented data to produce a plurality of data and parity blocks;
generating one or more indexes regarding the page of column oriented data;
generating statistical data regarding the page of column oriented data; and
arranging the data and parity blocks, the one or more indexes, the statistical data, and a key column as the long term storage data.

### Claim 36 (depends on 35)

The computer readable memory of claim 35, wherein the second memory section further stores operational instructions that, when executed by the loader computing node, cause the loader computing node to:
include metadata regarding the data and parity blocks in the long term storage data.
