Database cleaning

Your file is not wrong. It is just written by five different people.

Three date formats, a city spelled four ways and an amount your spreadsheet quietly refuses to add up. Upload your file and we tell you what it has. Free, no account.

No account, nothing uploaded.

This is what we mean

Nine rows of a sales export. Every problem below is one we actually detect.

Before
DateClientCityAmount
15/03/2026Grupo IberiaMEXICO2.450,75
17-mar-26Textiles del SurMéxico1890
18/03/2026Grupo IberiaMéxico3200
19-mar-26comercial andinabogotan/d
20/03/2026Comercial Andina Bogotá1.120,00
17-mar-26Textiles del SurMéxico1890
2026-03-21Naviera PacíficoLima 4300
22-mar-26NAVIERA PACIFICOlima2.780,50
2026-03-23Andes SoftwareSantiago5100
SUM of Amount16,380.00

What the spreadsheet adds up today. It is missing 4,461.25 and shows no error at all.

Highlighted cells are the ones that changed. Fictional data, real problems.

«De la Torre» is one surname, not two

Every tool that splits a full name assumes one first name and one last name. Ours knows that De la Torre, Montes de Oca and Del Río are single surnames — and when a name genuinely is ambiguous, we ask you instead of guessing and corrupting your list.

One column

María Fernanda De la Torre

What other tools do

María · Fernanda De la Torre

Split correctly

María Fernanda · De la Torre

What we find

In plain words. You decide what gets fixed.

  • Dates written three different ways

    15/03/2026, 17-mar-26 and 2026-03-21 in one column. Nothing can sort or group them.

  • Repeated rows

    Identical rows counted twice. We also flag the nearly-identical ones without touching them.

  • Broken accents

    México, Bogot� — text saved wrong somewhere along the way.

  • The same thing spelled four ways

    MEXICO, México and mexico are three different cities to a spreadsheet.

  • Numbers stored as text

    2.450,75 looks like a number and adds up as zero. This is the one that changes your totals.

  • Mixed columns

    An n/d or a pendiente among the numbers, which turns the whole column into text.

  • Rows that are not data

    The title above the table, the blank rows inside it, the TOTAL line at the end.

  • Compound Spanish surnames

    De la Torre stays together. Ambiguous ones we hand back to you, one by one.

Your file never leaves your computer

Not a qualified claim, and not a promise about how we store things. There is nothing to store: this whole section runs in your browser.

  • Nothing is uploaded

    There is no server in this flow. Not one row of your file crosses the network.

  • No AI involved

    The cleaning is deterministic rules, not a model. Nothing about your file is sent to anyone to be read.

  • Nothing is stored

    Close the tab and it is gone, including from us — because we never had it.

The details are in our Privacy policy.

How it works

  1. 1

    Upload it

    CSV or Excel. It is read in your browser, like opening it in a spreadsheet.

  2. 2

    We tell you what it has

    A plain-language report of every problem found. Free, and no account needed.

  3. 3

    You choose what to fix

    You see exactly what changes before anything is applied. Nothing is decided for you.

  4. 4

    Download it clean

    Same format in, same format out. Or send it straight into Dashboards without uploading again.

Find out what your file has. It costs nothing.

The diagnosis is free. Cleaning it costs 35 credits.

No account, nothing uploaded.

Let's see what is wrong with your file

Upload it and we will tell you, in plain words, what would make your numbers come out wrong. Nothing is changed and nothing is uploaded.

Drop your file here

or

Everything is read and analysed in your browser. No row of your file leaves your device — no server, no AI, nothing.

Suggestions, help or a problem? Email us at support@9raphi.app

We use cookies to understand how the app is used. Analytics stays off until you accept. Privacy policy