CSV Mixed Dates and Duplicates, No Time? The Badge Table

2026-11-02

It is 7:42 in the morning, and you are standing at a folding table in a hotel lobby.

The table holds a stack of blank name badges, a roll of tape and a lanyard box. Your laptop balances on a luggage trolley because there was nowhere else to put it. The coffee is the free kind, in a paper cup, already lukewarm.

The coach to the conference venue leaves at ten. Two hundred and twelve people are supposed to be on it. Someone at the venue needs the final arrivals list by nine, so the caterer, the room allocation and the badge printer all agree on who exists.

You open the export from the registration form. You scroll. Your stomach does the small drop you know well.

Three ways to write a Tuesday

The arrival column has 03/11/2026. It also has 2026-11-03. A few rows say 3 Nov. One says Tues 3rd, and you have no idea which month the person had in mind.

Is 04/11/2026 the fourth of November or the eleventh of April? For most rows it does not matter. For some it decides whether a person gets a room tonight or tomorrow.

Then you notice the names. Marta Kowalski is there twice. Once with a company email, once with a personal one. Below her, "J. Okafor" and "Jide Okafor" sit three rows apart with the same phone number. You start counting duplicates by eye, and you lose count at eleven.

The thought arrives, quiet and familiar: I should have fixed this on Friday.

But on Friday the form was still open. People kept registering. Everything felt too early to clean, and now it is suddenly too late.

The cost nobody puts on the invoice

Nobody sees this part of the job. The client sees a smooth morning, a full coach and badges that spell names correctly. They do not see you at a trolley, squinting at a column of dates and wondering whether to trust your own guess.

The real cost is not the file. It is the small, grinding fear that one wrong row becomes a visible mistake. A guest without a room. A caterer preparing for fourteen people who are really seven. A badge with a name spelled two ways, handed over with a smile that does not feel quite honest.

It also costs something you can't bill. You were going to spend this half hour greeting people, not auditing a spreadsheet. You wanted to look like someone who has it handled. Instead you are hunched over a laptop, sorting by column, hoping the sort doesn't make things worse.

You try the usual rescue. You sort by name, delete what looks like a duplicate, then wonder whether you deleted the right one, the one with the newer email, the one with the dietary note. You can't undo your way back through twenty minutes of edits. The file is changing under your hands, and you can't say what changed.

The pause

Then a colleague, Lena, walks past with a box of lanyards. She runs the registration desk and has been doing this for years. She glances at your screen, then at your face.

"Is that the export?" she says. "Don't touch the original."

That is all she says. She goes to find tape.

But the sentence stays with you. Don't touch the original. It's an obvious idea. Keep the raw file as it is. Let something else do the tidying. Look at what would change before it changes. Get a short report of what was merged, what was reformatted, and what was left for you to decide.

Not a heroic all-nighter, and not forty careful minutes of squinting. Just a calm pass over the whole file, with the changes laid out where you can read them.

It does not have to be like this. You just never had a routine for it. Every export was a fresh emergency, so every time you improvised.

The same lobby, a different Monday

Picture the other version of this morning.

You arrive at 7:30. The export has already been through a cleaning step, one you run the same way each time. Dates all read the same. Doubled names are merged, and the few unclear ones are flagged in a short list instead of silently guessed. A change report sits next to the file, and the original is untouched in its folder.

You spend ten minutes with that report. Two rows need a real human decision, so you send Jide a quick message. Then you put the laptop on the table, not the trolley, and start handing out badges.

At 8:50 the final list goes to the venue. You still have time to eat something. The coach pulls out at ten, and you watch it go with a feeling you rarely get at that hour: nothing behind you is waiting to blow up.

The work was the same. The feeling was not. You were not bracing for the next surprise, because you could see exactly what had happened to the data.

A gentle next step

If you recognise yourself in the lobby, or in whichever room your own version of this happens, you might want a routine instead of a rescue.

Clean Data in Minutes – CSV & Excel Cleaning Toolkit (Python) is built for this kind of morning. It cleans messy CSV and Excel exports in minutes: dedupe, fix dates, phones and encoding, split, merge and validate. It runs as a dry-run first and gives you a change report, so you see what would happen before anything is touched. It comes with a 30-day money-back promise: if it doesn't help, you get a full refund, no questions.

No need to decide today. Just keep the sentence from the lobby: don't touch the original.

An illustrative story, written with AI assistance.