Skip to content
DataUntangler
Buy the plan

One file free to try, no card. Already paid? Sign in.

Data scrubbing software and tools with a change log you can export

Data scrubbing software finds the defects that make a file untrustworthy, duplicate records, mixed date formats, broken characters, numbers stored as text, states written six ways, and corrects them. DataUntangler runs that scrub in your browser and shows the result as a row by row diff. You approve or reject each proposed scrub, and what you decided leaves with you as a CSV audit log.

In-browser your file never leaves your computer

Every row proposed to you before it is applied

CSV log of what you approved and rejected

Clean a file
01 Try it on this page

Load a messy sample and watch it work

The same engine described below, running on your own machine. Nothing is uploaded and nothing is applied until you approve it.

1. Load a file

Your file is read on your own machine. Nothing is uploaded, so nothing can leak.

Or start from a messy sample

Status

Loading the sample file.

File: CRM contact export

Rows

--

Columns

--

Changes

--

Duplicate rows

--

Date formats

--

Blank or invalid emails

--

Load a file to see the diff

Drop a CSV or pick one of the messy samples. Every proposed change appears here as a line you approve or reject.

02 About data scrubbing software

What a scrub actually takes out of a file

Most files arrive damaged in the same handful of ways, because the damage comes from the same places every time: a person typing at a trade show, an export that guessed at an encoding, a spreadsheet that helpfully turned a ZIP code into a number. A scrub is the pass that puts all of that back into one shape.

These are the rules this engine applies, one column at a time, after it works out what each column holds:

Defect How it shows up in the raw file What the scrub proposes
Encoding damage Café Bellini, á, ’ where a letter should be The original character restored
Whitespace Leading, trailing and doubled spaces, non breaking spaces One space between words, none at the ends
Casing acme corp, ACME CORP, AcMe CoRp Title case, with known acronyms left alone
Email format JANE.DOE@Acme.COM with a trailing space jane.doe@acme.com
Date format 03/04/2026, 4 Mar 2026, 2026-04-03 in one column One ISO date per row
Currency format $1,240.00 and 1240 in the same column A plain number a spreadsheet can total
Number format Digits stored as text, stray thousands separators A real number
Leading zero A Boston ZIP code that arrived as 2108 02108
State standard mass, Massachusetts, Mass. MA
Country standard U.S.A., USA, us United States
Duplicate records Two rows for one customer, spelled differently A scored cluster with the row to keep marked

Each proposal carries the rule that fired and a confidence label, so a scrub you disagree with is one click to reject rather than something to unpick afterwards.

Database scrubbing and scrubbing the export are two different jobs

Database scrubbing usually means a scheduled job that runs against a warehouse or an operational system, written once by an engineer and maintained forever. It is the right answer when the same bad data arrives every night from the same pipeline. It is a project, it needs a connection, and it needs somebody who can be asked to change a rule.

The far more common case is one file, on one laptop, that has to be right before a meeting. A migration extract. A conference list about to be imported. A finance export that has to reconcile. Nobody is going to schedule a warehouse job for that, and nobody should have to file a ticket to fix a date column. That is the job this tool does, and it is the same job covered on the wider data cleaning software page from the other direction.

If your scrub is really about the duplicate problem rather than the formatting, the matching side of it lives on data deduplication tools and fuzzy matching. If the scrub is part of a system move, the week by week cutover plan is in data migration cleansing software, and if you are weighing an enterprise matching platform for it, the Data Ladder alternative page says honestly where that kind of product is the better buy.

Scrubbing software, a scrubbing service, or a script you maintain

There are three honest ways to get a file scrubbed, and the right one depends less on your budget than on how often you have to do it and who has to defend the result.

Approach What you are buying Where it stops
A scrubbing service Somebody else does the labor, usually quoted per record or per project Your customer list leaves your building, so the agreement comes before the work, and the next file starts the quote again
A Python or SQL script Exact control and no licence cost Somebody has to own it, and the cases it misses are invisible until a number is challenged
An enterprise data quality platform Scheduling, connectors, governance, a whole practice Procurement, onboarding measured in weeks, and far more product than one file needs
Browser scrubbing software The scrub, the review and the log, today, on the file you already have One file at a time, and the browser demo is capped at 5MB

Outsourcing has a place. A one time cleanup across several source systems, with enrichment against outside reference data, is labor, and buying labor is reasonable.

We wrote the longer version of that comparison on data cleaning services, including the questions worth asking a vendor before the list leaves your network.

How a scrub runs on the file you have open right now

  1. 1 Drop the CSV or TSV onto the demo above. Parsing happens in your browser, so the file does not go anywhere.
  2. 2 The engine infers what each column holds: email, date, currency, ZIP code, state, country, name, company. The right rule then runs on the right column instead of on everything.
  3. 3 You get a list of proposed scrubs. Each one shows the original value, the new value, every rule that fired and how confident the engine is.
  4. 4 Duplicates come back as clusters rather than deletions, with the most complete row marked as the one to keep and the reason the rows matched attached.
  5. 5 Approve what is right, reject what is not, then download two files: the scrubbed CSV and the audit log of every decision.

Nothing is applied while you are reading. A scrub you reject never touches the output, which is the difference between a tool that cleans your file and a tool that changes it.

Why the log matters more than the scrub

Any tool can lowercase a column. The hard part comes three weeks later, when somebody asks why the March total moved and the only answer available is that the data was cleaned.

Because every scrub here is approved by a person and written into an exportable log, you can hand over the scrubbed file and the list of what changed inside it, keyed by row and column. That is what makes the output usable in a migration, an audit or a board pack. Keeping a file in that state rather than scrubbing it again every quarter is data hygiene, and the same review step is what separates this from the free desktop route described on the OpenRefine alternative page.

03 Read it off a real file

A real scrub, run on a four row sample

Column Before After Rule Confidence
Customer acme corp Acme Corp Whitespace and Casing Likely
Customer Café Bellini Café Bellini Encoding repair High
Email JANE.DOE@Acme.COM jane.doe@acme.com Whitespace and Email format High
Signup 03/04/2026 2026-03-04 Date format Likely
Signup 4 Mar 2026 2026-03-04 Date format High
Amount $1,240.00 1240.00 Currency format High
ZIP 2108 02108 Leading zero High
State mass MA State standard High
Country U.S.A. United States Country standard High

Every one of these came out of the engine on this page, not out of a slide. Note the two date rows: the same column held two formats, so a sort on it was meaningless until the scrub ran.

04 Scrubbing with find and replace

What a find and replace scrub leaves in the file

The usual scrub is a sequence of find and replace passes down a spreadsheet, run by whoever drew the short straw, and it fails in three predictable ways.

Excel data cleaning

It only fixes the spellings you thought of

You replace Mass with MA and Massachusetts with MA. The row that says Mass. with a full stop, and the one with a trailing space, both survive and neither is visible when you scroll.

It changes cells you did not mean

Replacing USA with United States also hits a company called USA Freight. Find and replace has no idea which column it is in or what that column holds.

It leaves no trace

When the totals move, there is no record of which cells were touched, in what order, by whom. The file is different and the reason is gone.

What a column aware scrub does instead

It works out what the column holds first, applies only the rule that belongs there, proposes every change with its reason, and writes your decisions to a CSV you can attach to the file.

05 Before you run it

Questions about data scrubbing software

What is the difference between data scrubbing and data cleansing?
In practice they name the same work. Scrubbing tends to be used for a one time pass over an extract before a migration or a consolidation, and cleansing for the ongoing discipline. The rules that run are identical: repair, standardize, validate, deduplicate.
How much does data scrubbing software cost?
DataUntangler is 59 dollars a month, or 348 dollars a year, for unlimited files up to 25MB each. Enterprise platforms in this category are quoted per seat or per record through procurement. The browser demo on this page needs no card and no account.
Can you scrub data without uploading the file?
Yes. The demo on this page parses and scrubs entirely in your browser, so the file never reaches a server. That is what makes it usable on a customer list you are not allowed to send to a vendor, and why there is no agreement to sign before you try it.
Does data scrubbing delete records?
Not here. Duplicates are proposed as clusters with a suggested row to keep, and nothing is removed until you approve that cluster. Rejected proposals never reach the output file, and the export includes a log of both.
How large a file can data scrubbing software handle?
The browser demo takes about 20,000 rows, or 5MB. The paid plan takes files up to 25MB, roughly 100,000 rows, with no limit on how many files you run. Beyond that you are into warehouse territory and a scheduled job is the better answer.
06 What it costs

Plans for when the file is bigger than the demo

Yearly billing saves 50 percent. There is no free plan, and no card is needed to use the browser demo above.

Yearly billing saves 50 percent, half the monthly price.

Switch to yearly and save 50 percent.

Analyst

Self-serve

One person with files to clean, as many as they like.

$29$59/mo

Billed $348 a year Billed monthly

Buy Analyst, yearly Buy Analyst, monthly
  • Unlimited files
  • Files up to 25MB, about 100,000 rows
  • Exact and fuzzy duplicate detection
  • Dates, numbers, currencies, encoding and casing fixed
  • Every change reviewable before you download
  • Cleaned CSV and a change log, both downloadable
  • Runs in your browser: files never leave your machine
  • Email support

Enterprise

Several people on one invoice, a DPA on request, priced per seat.

$49$98/seat/mo

Billed $588 a seat a year, from 2 seats ($1,176) Billed monthly per seat, from 2 seats ($196)

Buy Enterprise, yearly Buy Enterprise, monthly
  • Everything in Analyst
  • Several people on one invoice
  • DPA on request
  • Invoicing and purchase orders
  • A named contact for support

Payment is by card through our payment provider, monthly or yearly, and you can cancel from your account at any time. The free try is one file of your own, up to 5MB, in the browser demo. Already paid? Sign in.

Related pages

Clean the file, and be able to show what you changed

Try the demo on your own CSV first. It runs in your browser, it costs nothing, and no card is required.

No card required. Your file never leaves your computer.

Clean a file