> For the complete documentation index, see [llms.txt](https://v2.dataos.info/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://v2.dataos.info/build/understand/explore-data/datasets.md).

# Datasets

Inspect dataset metadata, lineage, and usage in the catalog.

The Datasets view is where you inspect a source's metadata before you reuse or productize it. It tells you what data exists, where it lives, how it relates to other assets, and whether it is trustworthy enough for your use case. Understand the dataset before you build on it.

Inspecting dataset metadata lets you:

* Find available datasets across connected sources.
* Read ownership, tags, descriptions, and column types.
* Check whether a dataset is already part of a Data Product.
* Inspect table-level and column-level lineage.
* Review recent query activity.

## Navigate

```
DataOS Home -> Datasets -> Select source -> Search or filter -> Open dataset
```

## 1. Open Datasets and select a source

From the left navigation, open **Datasets**. The source tree lists connected systems: Snowflake, PostgreSQL, Iceberg, custom databases, and more. Select the system where your dataset lives. The dataset list updates to show datasets from that source. See [Metadata sources](https://v2.dataos.info/references/resources/nilus/metadata-pipelines/metadata-sources) in References for how each source type is modeled in the catalog.

<figure><img src="https://797497916-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fo5cTluI9fPU0SarnfrVG%2Fuploads%2Fgit-blob-0b26c724ffe06af8ef2b88271e90a426bca25ddf%2Fimage%20(2).png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
The dataset list can include datasets ingested by other Tenants on the same Instance. Ingested dataset metadata is visible across the Instance so teams can discover what raw material is available — but visibility is not access. Opening a dataset here does not grant the Depot, Secret, or access policy needed to read the underlying data.
{% endhint %}

## 2. Search and open

Search by name, schema, or keyword. You can hide or show datasets already productized. In the list, scan the name, source path, tags, owner, and the productized indicator to decide what is worth opening, then open it.

The detail page shows the name, source path, and productization status:

```
CUSTOMER
Productized into 1 data product
```

Inside a dataset, three tabs:

| Tab      | What it shows                                                            |
| -------- | ------------------------------------------------------------------------ |
| Overview | Columns, tags, descriptions, version, update time, productization status |
| Lineage  | Upstream and downstream flow at table or column level                    |
| Queries  | Recent query activity                                                    |

{% hint style="info" %}
If a dataset is already productized, inspect that Data Product before building a new one. It may already solve your need.
{% endhint %}

## 3. Inspect the Overview

Use the Overview tab to read the structure:

| Check               | Why it matters                                 |
| ------------------- | ---------------------------------------------- |
| Column names        | Confirms the fields you need are present       |
| Column descriptions | Explains business meaning                      |
| Tags                | Show role, sensitivity, domain, classification |
| Last updated        | Tells you whether the metadata is current      |
| Productized badge   | Shows whether it already powers a Data Product |

Then go column by column to answer: *can I safely use this column in a model, metric, or contract?* A primary-key indicator means it can be a join key; a data type tells you what transformations and validation you will need; a lineage icon means you can trace the column further.

## 4. Inspect Lineage

Open the **Lineage** tab to see how the dataset connects to other assets: source systems, base tables, productized datasets, downstream models, semantic assets, and column-level dependencies.

Use the **Columns** toggle to switch between table-level and column-level lineage. Table-level gives you the high-level flow; column-level tells you which input columns feed an output, how a field carries through transforms, and which downstream outputs a change to that column would affect. This is the difference between general discovery and precise impact analysis.

Use lineage before relying on a dataset as an input: it shows whether changes to it can break downstream products.

## 5. Review Queries

Open the **Queries** tab to see recent activity: submission time, who ran the query, and frequency. This tells you whether the dataset is actively used and who has working knowledge of it, both useful signals of whether it is safe to depend on.

## Decide what to do next

| Decision                       | Choose it when                                                                                                                                       |
| ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
| Productize the dataset         | It has the fields, quality, and lineage your use case needs                                                                                          |
| Explore the values             | The metadata looks right but samples need validation. Open it in [Workbench](https://v2.dataos.info/build/understand/explore-data/workbench.md).     |
| Inspect a related Data Product | It is already productized and may solve the need                                                                                                     |
| Contact the owner              | Meaning, access, or reliability is unclear — including when a dataset was ingested by another Tenant and you need your own Depot or Secret to use it |
| Move data in                   | The data you need is missing. See [Move](https://v2.dataos.info/build/move/overview.md).                                                             |
| Avoid the dataset              | It is incomplete, stale, unowned, or unsuitable                                                                                                      |

When the metadata looks promising, validate the actual values in [Workbench](https://v2.dataos.info/build/understand/explore-data/workbench.md) before you commit.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://v2.dataos.info/build/understand/explore-data/datasets.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
