Excel Turned My Product Code 1E5 Into 100000 — It Was Never a Number

Excel Turned My Product Code 1E5 Into 100000 — It Was Never a Number

Don't double-click the supplier CSV. The moment a spreadsheet opens it, codes like 1E5 stop being codes — they are read as numbers, and if you save, the change is permanent. Check the file first instead: run it through the checker, download the value-preserving XLSX, and 1E5 stays the text 1E5.

Try the toolPrevent Excel Data Loss: CSV Value CheckCheck CSV and Excel data for values spreadsheet apps might change, then export an XLSX that preserves them.loopyti.com

One warning before you start: if you already opened the file, watched the codes turn into numbers, and saved it, the original codes are gone. The tool can't bring back information that no longer exists — it can only protect a file that still has its original values. If you still have the original CSV, check that one.

![Rendered view of the supplier price list as a spreadsheet shows it after a normal open, with product code 1E5 shown as 100000](/blog/media/2026/11/6ff15c08-82b6-486d-9595-abe557b7365b.png

How to keep the codes in three steps

  1. Add the CSV to the checker. Upload the file — inspection runs in your browser, so the file is never sent to a server. The tool detects the encoding (mine came back UTF-8) and checks every cell against the patterns spreadsheet apps misread.
  2. Read the results. The checker reports how many cells face each risk and where they are. My file came back "Check complete" with 8 at-risk cells, all of them "May be interpreted as scientific notation" — rows 2 to 9, column 1, the eight codes in order: 1E5, 2E7, 3E4, 9E1, 5E2, 7E3, 4E6, 6E8. The summary also counts "Overlapping code groups 2" — the collisions, listed in their own section below the preview. Each one carries the recovery status "Can be corrected" and a recommended output identical to the original.
  3. Export the value-preserving XLSX and open it. Click "Download value-preserving XLSX · Recommended" — a toast appears: "Value-preserving XLSX saved." The at-risk values are stored as text, so a spreadsheet leaves them alone.

![CSV checker results showing the new Overlapping codes section with two groups: 1E5 and 100000 both becoming 100000, and 5E2 and 500 both becoming 500](/blog/media/2026/11/9be732ac-7ca1-4a92-90f8-3eb97bed8de6.png

Does it actually work? I tested it

I made a fictional supplier price list — "Maple & Thread Supply Co.", 24 rows, four columns: product code, product, unit price, quantity. Eight of the codes look like scientific notation: 1E5, 2E7, 3E4, 9E1, 5E2, 7E3, 4E6, 6E8. Two more codes are plain numbers, 100000 and 500. The other fourteen are hyphenated codes like CT-104 and HB-2210. Tested 2026-11-13, and re-tested the same day after the tool was fixed (the first version of this post described the old behaviour — see below).

Open that CSV the normal way and a spreadsheet rewrites 8 of the 24 codes. I checked the same file in LibreOffice first, which produced exactly these values:

Code as written What the spreadsheet shows
1E5 100000
2E7 20000000
3E4 30000
9E1 90
5E2 500
7E3 7000
4E6 4000000
6E8 600000000

The worst part isn't the new numbers — it's the collisions. My list already has a code 100000 (a cream ribbon spool, same product family as the sage spool coded 1E5) and a code 500 (a butter-yellow felt sheet, next to the dusty-pink sheet coded 5E2). After a spreadsheet opens the file, 1E5 and 100000 are both 100000, and 5E2 and 500 are both 500. Two different products now share one code, and nothing in the sheet says which row is which. Twenty-four distinct codes go in; twenty-two distinct values come out.

![Table comparing product codes as written in the CSV with what the spreadsheet shows, with 1E5 changed to 100000](/blog/media/2026/11/d0ab3af0-a7d1-4138-b13e-c55d60291c5f.png

The fourteen hyphenated codes are untouched — CT-104 stays CT-104 — because they don't look like numbers. Only the codes shaped like numbers get rewritten.

The checker caught exactly the eight. It flagged all eight scientific-notation codes and nothing else: the plain-number codes 100000 and 500 were not flagged, and neither were the hyphenated codes. That is the right call per cell — 100000 and 500 are numbers already, so there is nothing for the checker to warn about in isolation. But this time the checker also names the collisions: a new "Overlapping codes" section says "Different codes that Excel turns into the same number can no longer be told apart. Saving the value-preserving XLSX keeps the originals as they are", and lists both groups — "Value shown in Excel 100000" with 1E5 · Row 2 and 100000 · Row 10, and "Value shown in Excel 500" with 5E2 · Row 6 and 500 · Row 11. So the clash is no longer yours to spot alone: the result screen shows exactly which two codes will become one value.

The export held up. The downloaded file, supplier-price-list-text-values.xlsx (20,856 bytes), reopened with all 24 rows intact, and I checked every cell: all eight codes came back as text — 1E5, 2E7, 3E4, 9E1, 5E2, 7E3, 4E6, 6E8, exactly as written. The plain codes 100000 and 500 came back as text too, and so did every price and quantity — the export stores the whole sheet as text, not just the flagged cells.

![Value-preserving XLSX reopened showing product code 1E5 unchanged as text](/blog/media/2026/11/c30d5f10-1d27-4bf1-a19d-97b97f19b157.png

Why a spreadsheet does this

A CSV has no types. It is plain text — there is nothing in the file that says "this cell is a code, that cell is a price." So the spreadsheet guesses the type of every cell from its shape when it opens the file, and 1E5 has the shape of scientific notation: 1 × 10⁵. The code was a name a supplier made up; the spreadsheet sees the shape and calls it a number. That is the whole mechanism, and it is also why CT-104 survives: a hyphen makes the shape unmistakably not-a-number, so it is left alone.

What about the other ways?

Open it through the import wizard and set the column to Text works — but you have to remember to do it every single time a supplier sends a file, and one double-click on the wrong day ruins that file. It is a habit, not a fix.

Import it into Google Sheets instead doesn't help either. With default settings, Sheets guesses types the same way — the checker's own example page shows a Sheets import changing the values. The guessing happens wherever a CSV is opened, not only in Excel.

Put an apostrophe in front of each code in the CSV works, but you are editing the data to protect it, and the supplier's next file arrives unedited anyway. You would be patching every delivery by hand.

The checker's place is the case before any of that: a file you haven't opened yet, where you want the original values kept as text without touching the data yourself. It stops the change from happening; it doesn't undo one that already happened.

What it won't do — read this first

  • It can't restore values that were already damaged. Once a code was saved as a number, the original text is gone. Each flagged cell in my test carried the status "Can be corrected" — that is the checker's verdict for values that still exist in the file, not a promise about values already lost.
  • It now warns about collisions. Two different codes that land on the same number — 1E5 and 100000, 5E2 and 500 — are listed in an "Overlapping codes" section with the rows each one came from. (An earlier version of this post said the checker never mentioned them — that was true before the fix, and it is fixed now.)
  • No XLS files, and size limits apply. CSV and XLSX only; files over 10 MB or 50,000 rows are rejected, and only the first sheet is inspected and exported.
  • Take the XLSX, not the CSV, if you want the codes kept. The checker also offers a UTF-8 BOM CSV download, but that is for preventing garbled text — it does not guarantee leading zeros, dates, or long numbers stay preserved. Only the XLSX stores the values as strings.
  • A clean result means "no risk signals found", not "the data is correct." The tool doesn't check for typos or wrong values.
  • Your file stays on your machine. Inspection and export run in your browser; nothing is sent to a server, so a customer list stays a customer list.
  • This post is only about the short-code case. Long numbers losing their last digit, postal codes losing a leading zero, and dates like 3-4 turning into March 4 are different symptoms of the same tool — they are covered in Why Excel Breaks Your CSV: Long Numbers, Leading Zeros and Dates.

The one line to remember

Check before you click — 1E5 is a code, not a hundred thousand.

Try the toolPrevent Excel Data Loss: CSV Value CheckCheck CSV and Excel data for values spreadsheet apps might change, then export an XLSX that preserves them.loopyti.com

FAQ

Can the tool fix codes a spreadsheet already converted? No. Once the original value was saved as a number, there is no information left to restore it accurately. Review the suspected damage in the results, and if you still have the original CSV, check that file instead.

Does a clean result mean the file is safe? It means the inspection rules found no risk signals. It doesn't check for typos or incorrect data, so a code that was wrong in the source stays wrong.