Skip to main content

Flow Insiders Letter No.3

This is the No.3 letter shared with Columns Flow builders.

Thanks for being part of the early access group. This is our third update, and it reflects feedback from real usage.

Since our last check-in, I have talked with several of you about your use cases. It is great to see Columns Flow used in production-like work, and I appreciate the honest feedback about bugs and blockers.

Please keep the feedback coming. It is exactly what we need right now.

The most frequent question has been: Can I work with multiple inputs at once?

Yes. Union, Join, and Compare operations are now supported. You can merge multiple data sources and run cross-input analysis in one flow.

Quick intro to multi-input operations:

Flow-Join

In this example, there are three node types:

  • Green nodes: input nodes where you connect sources
  • Blue nodes: transform nodes where you define transformations
  • Yellow nodes: merge nodes that combine multiple inputs

Here is a walkthrough of one example:

Problem: I have two sheets in this Google Sheet: Orders and Customers.

Goal: Get total sales by customer first name.

orders.png customers.png

becomes

amount_by_first_name.png

note

Let us do this in Columns Flow.

Steps:

  1. Create a flow and connect the default input node to the Orders sheet.
  2. In the Flow Info panel, click +Input Node, then connect the new input node to the Customers sheet.
  3. Add a transform node to extract first names from the customer name field. Generate and approve the step so first_name is available.
  4. Fork a new node and connect both Orders and transformed Customers to it. It becomes a merge node, and merge options will appear.
  5. Add a node after the join and type total amount by first name as the next transformation. Turn on graph view to confirm the result.

In step 4, choose one of these merge modes:

  • Union: concatenate rows from multiple sources
  • Join: keep only matching rows on join keys (SQL Inner Join)
  • Lookup: keep all rows from the primary source and match others (SQL Left Join)
  • Except: keep rows in the primary source that have no matches (SQL Anti Join)
tip

You can schedule this flow to run automatically and configure alerts on any node.

Multi-input operations unlock much richer automation in Columns Flow.