Rename Columns

Transform

Rename one or more columns by declaring from → to mappings. Columns not listed in the mappings pass through with their original names. Use it to clean up source headers (e.g. Customer ID → customer_id), align columns to a downstream schema, or apply consistent naming across imported files.

Local processing

This transform runs entirely in your browser. No data is sent to external servers.

How it works

Rename Columns is a streaming transform: each row is processed independently and emitted with the matching keys swapped to their target names. Row order, row count, and cell values are unchanged — only the keys differ.

The transform builds a Map<from, to> once at startup and rewrites each row’s keys in input order. Columns whose names are not in the mapping are emitted unchanged. Mappings whose from column does not exist on a row are silently ignored for that row.

If no mappings are defined, all rows pass through unchanged.

Input: One tabular data connection. Output: The same rows with the configured columns renamed; other columns untouched.

Options

Mappings

An ordered list of { from, to } pairs. Both fields are required strings; either being empty is a configuration error.

Field	Type	Description
from	string	Existing column name to rename. Required, non-empty.
to	string	New column name. Required, non-empty.

Defaults to an empty list (passthrough).

Examples

Normalize CSV headers to snake_case

The source export uses Title Case headers and you want lowercase snake_case for downstream tooling.

Before:

Customer ID	Full Name	Signup Date
1042	Alice Anderson	2024-03-12
1043	Bob Brown	2024-04-01
1044	Carol Chen	2025-01-18

Configuration: Three mappings — Customer ID → customer_id, Full Name → full_name, Signup Date → signup_date.

After:

customer_id	full_name	signup_date
1042	Alice Anderson	2024-03-12
1043	Bob Brown	2024-04-01
1044	Carol Chen	2025-01-18

Rename one column, leave others alone

Only the email_address column needs to become email; everything else stays as-is.

Before:

name	email_address	role
Alice	[email protected]	admin
Bob	[email protected]	user

Configuration: One mapping — email_address → email.

After:

name	email	role
Alice	[email protected]	admin
Bob	[email protected]	user

Tips and Edge Cases

• Renaming to a name that already exists collapses the two columns. The transform writes to the target key as it walks the input row in insertion order, so if you rename legacy_total → total while a total column already exists, the column written last wins. This is rarely what you want — pair with Select Columns to drop the duplicate first, or rename the existing total column to something else as a separate mapping. See apps/web/src/transforms/rename-columns/logic.ts:37-44.
• Mappings whose source column doesn’t exist on a row are silently ignored. This is intentional — sparse rows (where a column may be absent on some rows) don’t error. The downside is that typos in the from field don’t produce a runtime warning; the rename just never fires.

Related Transforms

• Select Columns — drop the leftover columns after a rename, or remove the duplicate when renaming into an existing name.
• Reorder Columns — change column order without changing names.
• Formula — derive a new column from existing ones if you need more than a name change.