---
title: "Anatomy of a dataset"
slug: "anatomy-of-a-dataset"
updated: 2026-01-28T10:32:21Z
published: 2026-01-28T10:32:21Z
canonical: "support.lusid.com/anatomy-of-a-dataset"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://support.lusid.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Anatomy of a dataset

A **dataset** is equivalent to a relational database table.

To get started, you [define the table schema](/v1/docs/how-do-i-create-a-relational-dataset-definition) in a **relational dataset definition**. Note the primary key consists of multiple columns, some nominated by LUSID and some by you.

You can then [add records to the dataset](/v1/docs/how-do-i-add-records-to-a-custom-dataset). A record with unique values in the primary key constitutes a **data series**:

![](https://cdn.document360.io/d575ad81-c0ed-4980-bbd1-d59ac5c3de82/Images/Documentation/DatasetAnatomy1 (1).png)

Adding a record with a different value in any column in the primary key constitutes a new data series (example cell change highlighted in red below):

![](https://cdn.document360.io/d575ad81-c0ed-4980-bbd1-d59ac5c3de82/Images/Documentation/DatasetAnatomy2 (1).png)

Adding a record with the same primary key but a change in the date column constitutes a new **data point** within an existing data series:

![](https://cdn.document360.io/d575ad81-c0ed-4980-bbd1-d59ac5c3de82/Images/Documentation/CustomDatasets3.png)

Adding a record with just a change in a value or metadata column updates an existing data point:

![](https://cdn.document360.io/d575ad81-c0ed-4980-bbd1-d59ac5c3de82/Images/Documentation/CustomDatasets4.png)

By default, [querying a dataset](/v1/docs/how-do-i-query-a-custom-dataset) retrieves data points within data series. You can optionally choose to just retrieve the data series.
