UnknownFebruary 25, 2026 • 4 min read

History Tables

Weld now supports history tables, allowing you to track changes in your data over time. Learn how to set up history tables in Weld and use them to gain insights into data trends and changes.

by Gabija Makutėnaitė

History Tables: A New Way to Track Changes in Your Data

Tracking the Past: Slowly Changing Dimensions (SCD)

Introduction: Why Tracking Historical Changes Matters

Imagine you’re analyzing your customer data and want to know:

Where did a customer live last year?
What was a product’s price in January?
Which salesperson managed a specific account in 2022?

If your database always updates records in place, you lose this crucial history. You can only see the current value missing out on trends, compliance reporting, and deep business insights. This is where history tables and the concept of slowly changing dimensions (SCD) help.

Slowly Changing Dimensions (SCD)

In data warehousing, a dimension is a way you describe objects in your business like customers, locations, or products. For many use cases, these dimensions change slowly but significantly: a company moves address, an employee changes department, a product gets rebranded, etc.

The challenge is: How do you handle changes in these dimensions over time?
This is where the Slowly Changing Dimensions (SCD) modeling technique comes in. SCD isn’t just one process, it’s a set of strategies for tracking changes in your data.

Types of SCD

SCD Type 1: Just Overwrite (No History)

The latest value simply replaces the old value.
Useful for non-critical changes.
No history is kept.

customer_id	address
123	SF

SCD Type 2: Full History Tracking

Each change creates a new row in a table (similar to change data capture).
Typically uses valid_from and valid_to columns to show when each version was true.
You can reconstruct the past!

customer_id	address	valid_from	valid_to
123	NY	2022-01-01	2023-08-10
123	SF	2023-08-10	9999-12-31

SCD Type 3: Limited History

Stores the current and previous value only (e.g., “current_address”, “previous_address”)
Fast to access, but you lose deep history.
customer_id current_address previous_address
123 SF NY

SCD Type 4: Separate Table

The main table keeps only current values, while a separate table with the same schema records all historical changes.
Often used when you want high performance for “current” analytics and full auditability.

SCD Hybrid Type and Beyond

Combinations of the above for different business needs.

Summary Table:

SCD Type	Keeps Full History?	How?	Example Use
1	No	Overwrite inplace	Non-critical
2	Yes	Row for each version	Deep audit
3	Partial	Current/Prev columns	Last-change
4	Yes (external)	Uses separate table	Compliance
6	Yes (hybrid)	Combos of the above	Analytics

History Tables

History tables are the backbone of SCD Type 4 implementations. It is a specialized table in your data warehouse that records every change made to your data.

Main table: Holds only the latest version of every record
History table: Keeps a timeline of all changes, so you can look up how data looked in the past

Example:
When a customer moves, the main table updates their address, and the previous address disappears.
But the history table adds a row for the change, so you can always reconstruct their address at any point in time. So, the schema in both tables is the same, but the history table has multiple rows for old records.

customer_id	address	updated_at
123	NY	2022-01-01
123	SF	2023-08-10

Using History Tables in Weld

Weld makes it easy to bring history tables into your analytics workflows for Shopify and Shopify Plus connectors.

How to add a history table with Weld:

Go to Data Source Configuration:
In the Weld interface, select your data source.
Search for History Tables:
Use the search panel and type history.
All tables ending with history in their name will be shown.
Add to Your Project:
Select the relevant history tables (e.g., customers_history, products_history) and add them to your pipeline. Weld will automatically bring in all the change-tracked data for those dimensions!