Datasets

A dataset is a named block of structured data — JSON, CSV, XML, key-value properties, or plain text — stored in one place and reusable across any job in Qinfinite. Where a global variable holds a single value, a dataset holds an entire data structure: a lookup table, a configuration block, a routing map, or a reference list.

Datasets also support versioning. Every time you need to update the content, you can create a new version rather than overwriting the existing one. Previous versions are preserved and can be reactivated if needed.

How It Works

Create v1

Save your first version of the data — it becomes active immediately

Create v2

When the data changes, create a new version — v2 activates, v1 is preserved

Switch Anytime

Activate any previous version if you need to roll back

Only one version is active at a time — jobs always use the active version.

Datasets vs. Global Variables

Both store reusable values, but they serve different needs:

Need	Best choice
One email address	Global Variable
A list of department email addresses	Dataset
One approval limit	Global Variable
Approval rules for every department, stored as JSON	Dataset
One API base URL	Global Variable
A lookup table of system names and their URLs	Dataset
A sensitive token or password	Global Variable (with encryption)
A multi-line configuration file	Dataset
A value that needs versioning	Dataset

Use a global variable for a single, self-contained value. Use a dataset when the reusable content is a structured block of data that may grow, change over time, or need a version history.

Dataset Types

When you create a dataset you choose a type. The type describes the format of the data you're storing. The built-in Monaco editor accepts any text, but choosing the right type allows Qinfinite to apply format-specific validation and beautification.

Type	What it's for	Typical example
JSON	Structured objects and arrays	Approval rules, configuration maps, lookup tables
CSV	Tabular data in comma-separated rows	Employee lists, routing tables, product codes
XML	Hierarchical markup data	Integration payloads, legacy system exports
Properties	Key-value pairs in key=value format	Application configuration, environment settings
Text	Any plain text content	Free-form notes, email templates, multi-line scripts

Creating a Dataset

1. Open Global Configurator and go to the Datasets tab.
2. Click Add Dataset.
3. Fill in the form:

Field	Rules	Notes
Name	4–30 characters, letters, numbers, underscore, or hyphen	Must be unique — checked as you type
Description	Optional	Helps teammates understand what this data is for
Tags	Optional, chip-style input	Useful for filtering and searching
Type	Required	JSON, CSV, XML, Properties, or Text
Value	Required	Entered in the Monaco code editor

4. The value editor is a full Monaco editor — the same editor used in VS Code. You can resize it by dragging the bottom edge. The default height is 300 px.
5. Use the Beautify button to auto-format your content (especially useful for JSON and XML).
6. Use the Validate button to check whether your content is well-formed for the selected type.
7. Click Create to save.

The first saved version is automatically labelled v1, set as the active version, and the total version count is set to 1.

Monaco editor shortcuts

The editor supports standard keyboard shortcuts: Ctrl+Z to undo, Ctrl+/ to toggle comments (for JSON-style content), and Ctrl+F to find within the editor. You can also drag the editor's bottom edge to make it taller when working with large data blocks.

Versioning

Versioning lets you update a dataset without losing the history of what it contained before.

Creating a New Version

In the dataset table, locate the row for an active version and click the Create Version action (shown as a git-merge icon). This:

1. Deactivates the current active version (it stays in the list but is no longer active).
2. Creates a new record with the next version number — for example v2 after v1.
3. Opens the edit form pre-filled with the previous version's content so you only need to change what's different.
4. Sets the new version as active on save.

Viewing Version History

Datasets with more than one version show a chevron (expand arrow) in the table row. Click it to expand the row and see all versions. Inactive versions are listed in descending version order. Each version shows its version label, who created it, and when.

Name is locked after versioning

Once a dataset has more than one version, the Name field is disabled in the edit form. The name cannot be changed after versioning begins because all versions share the same identity.

Activating an Older Version

To roll back to a previous version:

1. Expand the dataset row to view its version history.
2. Find the version you want to restore.
3. Click Activate on that version row.

The platform immediately deactivates all other versions for this dataset and marks the selected version as active. Any job that references this dataset will use the newly activated version at its next run.

One active version at a time

Only one version can be active at a time. Activating an older version deactivates the current one. Jobs always read the active version — there is no way to pin a job to a specific version number.

Referencing a Dataset in a Workflow

Once saved, a dataset can be referenced inside any job step that accepts dynamic expressions. The reference syntax is:

{{ds.datasetName}}

For example, if your dataset is named Dept_Approval_Rules, use:

{{ds.Dept_Approval_Rules}}

The platform substitutes the full content of the active version at runtime. Only active, non-deleted datasets are returned when a job looks up a dataset reference.

To copy the reference format quickly, click the copy icon next to the dataset name in the table.

Validation and Beautify

The editor footer provides two utility actions:

Button	What it does
Beautify	Auto-formats and indents the content based on the selected type. Particularly useful for minified JSON or XML.
Validate	Checks whether the content is valid for the selected type — for example, whether JSON parses correctly or whether XML is well-formed. Errors are shown inline.

Use Validate before saving if you're pasting in content from an external source. An invalid dataset saves successfully but may cause runtime errors when a job tries to parse the content.

The Dataset Table

The main table shows all datasets with these columns:

Column	Notes
Name	Chevron shown if multiple versions exist — click to expand
Type	JSON, CSV, XML, Properties, or Text
Value	Truncated preview of the active version's content
Version	Current active version label (v1, v2, etc.)
Tags	All tags for the dataset
Created	Date the first version was created
Last Modified	Date of the most recent change
Actions	Create Version (active rows only), Edit, Delete

Deleting a Dataset

Datasets use soft-delete — a deleted dataset is hidden from the table and excluded from search results, but its data is retained in the database. If a job references a deleted dataset, the reference will not resolve and the job step will fail.

Check references before deleting

Search for the dataset name across your jobs before deleting it. A deleted dataset cannot be referenced, and there is no automated warning sent to job owners.

Related Pages

• Global Variables — for single reusable values, secrets, and typed inputs
• Import and Export — move datasets between environments