Friendly Clean (fc) — Proposal

Name collision

fc is now a live command: filter-friendly-current (keyboard shortcut z Shift+C) filters the column under the cursor. See Filtering. Pick a different longname for Friendly Clean when implementing this proposal, for example fcl, or move it behind the Space palette with an unambiguous name.

A proposed batch data-cleaning command that chains multiple cleaning operations in one pass. Inspired by real data quality issues found in the practice files under 04-cleaning/.

The Problem

Real-world CSV data has recurring issues that require multiple manual steps:

Problem	Example	Source
Inconsistent casing	completed, COMPLETED, Completed	01-dirty_orders.csv
Negative values	-50.00 amount	01-dirty_orders.csv
Non-numeric in numeric column	abc in amount	01-dirty_orders.csv
Leading/trailing whitespace	common in CSVs	Various
Null/empty cells	blank region	01-dirty_orders.csv
Duplicate rows	Alice row repeated 3x	01-dirty_orders.csv
Zero values that should be null	0.00 amount	01-dirty_orders.csv
Missing emails	Carol, Grace, Kate	users.csv
Missing dates	blank last_login	users.csv

Proposed Syntax

Space fc Enter     # prompt: fc: <col>|* <op>[, <op> ...] [, <col> <op>...]

Operations

Operation	What it does	Applies to
lowercase	.lower() on text	text columns
uppercase	.upper() on text	text columns
trim	.strip() whitespace	all columns
abs	abs() on numeric	numeric columns
to-numeric	Cast to float, flag errors	numeric columns
fill-down	Copy non-null from previous row	any column
fill-missing	Replace empty cells with placeholder text	any column
dedupe	Remove duplicate rows (keeps first)	all *
zero-to-null	Replace 0 with empty	numeric columns
strip-punctuation	Remove ,.-_/ from text	text columns

Column targeting

Follows the same convention as fr:

Input	Target
status lowercase	Single column
* trim	All visible columns
amount abs to-numeric	Chained ops on same column
status lowercase, * trim	Comma separates clauses
status lowercase and amount abs	Cross-column via and

Real Examples

Example 1: Clean 01-dirty_orders.csv

Raw data:

amount    status     region
150.00    completed  sg
-50.00    completed  
200.00    COMPLETED  sg
abc       pending    jk
150.00    completed  sg
          shipped    my

Space fc → * trim, amount abs to-numeric, status lowercase, region fill-down, * dedupe

Result:

amount    status     region
150.00    completed  sg
50.00     completed  sg
200.00    completed  sg
#ERROR    pending    jk
150.00    completed  sg
0.00      shipped    my

Each step:

Step	Before → After
* trim	No visible change (whitespace stripped)
amount abs	-50.00 → 50.00
amount to-numeric	abc → #ERROR (flagged)
status lowercase	COMPLETED → completed
region fill-down	blanks filled from previous row
* dedupe	Alice duplicate removed

Example 2: Clean users.csv

Space fc → email fill-missing, last_login fill-missing, * trim

• Missing emails → placeholder no-email@unknown
• Missing last_login → placeholder never
• All fields trimmed

Why It Matters

Without fc:

• g* one pair at a time for text normalization
• g= with Python expressions for numeric transforms
• Manual navigation to each column
• No built-in abs, fill-down, or dedupe as one-step operations

With fc:

• Chain multiple cleaning operations in one command
• Same column-targeting convention as ff and fr
• Batch undo with U (same mechanism as fr)
• No Python syntax needed

Implementation Notes

• Build as keybind-source/38-friendly-clean.py
• Each operation is a simple Python function on the cell value
• Chained operations on same column apply in order (left to right)
• Uses the same vd.addUndo() batch-undo pattern as fr
• Follows same and/or / comma parsing as ff and fr

This is a proposal — not yet implemented. See the built commands ff (Filtering) and fr (Incremental & Replace) for the working pattern.