Most teams don’t work from one clean dataset.

Sales lives in one system. Marketing in another. Customer data somewhere else. None of it lines up by default.

Data blending lets you combine what you need, when you need it, without moving everything into a warehouse or setting up a full pipeline.

You connect sources, define the link, and the tool brings them together in a single view.

It’s not about building the perfect dataset. It’s about answering questions with what you already have.

‍

What Is Data Blending?

Data blending is a way to combine data from different sources into one analysis. You don’t merge the data in storage. You don’t create a new table.

Each source is queried separately. The results are then combined in the view based on a shared field like date, region, or product.

One source becomes the primary. Others are secondary. The primary defines what rows appear. The secondary fills in extra details where there’s a match.

Blending is useful when:

• You’re working across different platforms
• The data sources have different structures or levels of detail
• You need results quickly and can’t wait for a full integration

‍

How Data Blending Works

Blending happens at the view level, not in the backend.

You connect to two or more sources. The first field you drag into the view becomes the primary. When you drag a field from another source, it becomes the secondary.

If a shared field exists with the same name, the tool links them automatically. If not, you create the link manually.

Important rules to remember:

• Each source runs its own query
• Data is aggregated before blending
• You can’t mix raw and aggregated fields in the same formula

Because of this, secondary data is always aggregated. It only returns values that match something in the primary source.

For example, if the primary has only three months of data and the secondary has twelve, only the three shared months will appear in the view.

‍

When to Use Blending Instead of Joins

Joins combine tables before the analysis begins. They work when the data lives in one place and follows a consistent structure.

Blending works when that structure is missing.

‍

Use blending when you:

• Need to work across separate systems
• Want to use published sources
• Can’t or don’t want to change the data model

‍

Key Differences

• Joins happen before aggregation. Blends happen after.
• Joins create a new table. Blends don’t.
• Joins can include unmatched rows. Blends only show matches from the primary.

‍

Use joins when:

• Everything lives in the same database
• You need row-level operations
• You plan to reuse the joined dataset

‍

Use blends when:

• Data comes from different platforms
• You’re dealing with inconsistent structure
• The analysis is short-term or specific

‍

Choosing the Primary and Secondary Sources

Your choice of primary and secondary matters. The primary defines which rows appear.

The secondary only adds values where there’s a match.

‍

What to Watch For

• If a value is missing in the primary, the secondary won’t show it
• If the primary has three months and the secondary has twelve, you’ll only see three
• Switching the primary can reveal more data

Be intentional.

• Choose the primary based on your goal
• Know what might get excluded
• Test both directions to confirm results

‍

Working Across Multiple Data Sources

When you mix data sources, your options change.

Each calculated field must live in one source. If you reference a secondary field, it must be aggregated.

‍

What This Means

• Fields from a secondary source are always aggregated
• You can’t mix raw fields and aggregated fields
• You use dot notation to reference fields from another source (for example, [Sales].[Target])

‍

Limits to Expect

• You can’t sort by secondary fields
• Some filters won’t behave the same
• You may see an asterisk when the data doesn’t match cleanly

‍

To reduce issues:

• Keep most logic in one source
• Match field names where possible
• Use manual links when automatic ones fail

‍

Setting Up a Data Blend

Blending works per view, not across an entire dashboard or project.\

‍

Steps to Blend

1. Connect to your sources. They must appear separately in the data pane.
2. Add a field from one source. This defines the primary.
3. Add a field from another source. This becomes the secondary.
4. Confirm the relationship. If no link appears, define it manually.

If fields have different names, rename them or create a custom mapping.

You can blend on one field or several. But more links make the blend stricter and may hide rows.

Blending doesn’t create a new dataset. It builds a temporary link to support one view, for one question, using only the fields needed.

‍

Data Blending and Aggregation Rules

Blending always aggregates the secondary data. That’s how it works.

‍

What You Need to Know

• You can’t get row-level detail from the secondary
• Every field from it must be aggregated
• Even dimensions are wrapped in ATTR() by default

If there are multiple values for a field, Tableau uses an asterisk to flag the conflict.

‍

Example

You want to compare actual sales (primary) with quotas (secondary).

To get the ratio, you write:

SUM([Sales]) / SUM([Quota])



If you forget to aggregate both sides, the formula won’t run.

This matters most when combining real-time data with static summaries. You can only show totals or averages, not every transaction.

‍

FAQ

‍

What is data blending?

Data blending combines data from different sources into one view. Each source is queried on its own and merged at analysis time. No data is moved or stored together.

‍

How is data blending different from joining?

Joins happen before analysis and build one dataset. Blends keep sources separate, aggregate each one, and align them in the view.

‍

When should I use blending?

Use it when:

• Data lives in different systems
• You need a quick answer
• You don’t want to change your data model
• You’re working with published or real-time sources

‍

What’s the difference between primary and secondary?

The primary source defines which rows appear. The secondary only contributes where there’s a match in the primary.

‍

Why do values go missing from the secondary?

Only matches appear. If the primary is missing a value that exists in the secondary, it won’t show up unless you switch the primary.

‍

Can I blend on more than one field?

Yes. But more linking fields mean stricter rules. You could filter out more than intended.

‍

What if field names don’t match?

Rename one to match or manually map the relationship. What matters is that the values line up, not the names.

‍

Can I calculate using both sources?

Yes. But any field from the secondary must be aggregated.

‍

Why do I see an asterisk in the view?

There are multiple matching values in the secondary. Tableau doesn’t know which one to use, so it shows an asterisk.

‍

Can I save or publish a blend?

No. Blends only exist inside the view. For reuse, join the data or combine it elsewhere.

‍

What are the limits?

Blending can’t:

• Return row-level detail from the secondary
• Support all filters
• Handle non-additive metrics like MEDIAN or COUNTD consistently

‍

Will it slow down performance?

It might. Each source is queried separately. Use fewer fields or summary data to keep things fast.

‍

Which tools support blending?

Data blending is supported in Tableau, Looker Studio, Power BI (limited), and Alteryx. Each one handles it a little differently.

‍

Summary

Data blending lets you work across tools, sources, and formats without changing your entire pipeline. You keep data where it lives, query it separately, and blend results in one view.

It’s perfect for:

• Quick comparisons
• Published datasets
• Disconnected systems

It has limits. The secondary data is always aggregated. Some values might not appear. And you can’t save the blend for reuse.

But when speed matters, and structure isn’t perfect, data blending gives analysts a way to keep moving. It’s the shortcut that works when full integration isn’t worth the time.