Static Columns
Static columns store fixed values in a dataset that only change when manually updated.
About
A static column holds data that you provide directly. This includes inputs, expected outputs, labels, categories, or any fixed values. Unlike dynamic columns, static columns don’t run any computation. They only change when you update them manually or through the SDK.
For example, in this dataset the first three columns are static:
| user_query | expected_answer | category | model_response |
|---|---|---|---|
| What is the capital of France? | Paris | geography | (dynamic) |
| Summarize this article | A concise summary of… | summarization | (dynamic) |
You add user_query, expected_answer, and category yourself. The model_response column would be a dynamic column generated by running a prompt.
When to use
- Test inputs and expected outputs: Store the queries and ground truth answers for evaluation
- Labels and categories: Tag rows with classifications (e.g. “easy”, “hard”, “geography”, “math”)
- Default values: Pre-fill rows with consistent starting data when setting up a dataset
- Metadata: Store context like source, timestamp, or user ID alongside your test data
Supported Data Types
| Type | Description |
|---|---|
text | Strings and free-form text |
integer | Whole numbers |
float | Decimal numbers |
boolean | True or false |
array | Lists of values |
json | Structured JSON objects |
image | Image file references |
audio | Audio file references |
datetime | Date and time values |
How to Add a Static Column
You can add static columns through the UI or when creating a dataset via the SDK. See Add Columns to Dataset for step-by-step instructions.
Next Steps
- Dynamic Columns: Columns generated by prompts, evaluations, or models
- Add Columns: Add new columns to an existing dataset
- Create a Dataset: Start a new dataset from scratch
Was this page helpful?