DeskPlainOpen a blank workbook and use Data > From Text/CSV. Choose the separator that makes the preview correct, then use Transform Data to protect identifier columns as text before loading. Do not fix quoted CSV by replacing every comma.
If name;city;amount arrives in one cell, Excel probably used a different separator from the file. The file may be valid. Import a fresh copy rather than repeatedly saving the misread sheet over the original.
Fix it in about two minutes
- Keep the downloaded CSV unchanged and open a blank workbook.
- Choose Data > From Text/CSV and select the export.
- In the preview, try its actual delimiter: comma, semicolon, or tab.
- Choose Transform Data when columns need explicit types.
- Set identifiers such as ZIP codes or account references to Text before loading.
- Load into a new worksheet, check representative rows, and save the working result as XLSX.
Microsoft's text-import guide explains the preview delimiter control. The goal is not merely multiple columns: every value must land in the correct column without losing meaning.
| Preview symptom | Likely issue | Next check |
|---|---|---|
| Everything remains in one column | Wrong delimiter | Inspect the raw first few lines |
| An address becomes extra columns | Quoting or delimiter problem | Check text qualifiers and source export |
| 00124 becomes 124 | Identifier converted to number | Re-import the original as Text |
| Names show broken accented characters | Encoding mismatch | Choose the source's file origin/encoding |
Recognize separators and quoted values
Open a copy in a plain-text editor and inspect two or three complete records. For example, "Desk, large",2,49.00 has three comma-separated fields. The comma inside the quoted product name belongs to that value. A global find-and-replace would damage the record.
For "Desk, large";2;49,00, the semicolon separates fields while a comma may be the decimal mark. Check both delimiter and numeric interpretation. Do not split on spaces: a customer name or street address can legitimately contain them. Ask the exporter to repair malformed quoting rather than guessing at every ambiguous row.
Protect data types before Excel changes them
ZIP codes, product codes, and long identifiers are labels even when they contain digits. In Power Query's transform view, inspect any automatically added Changed Type step and assign Text before a numeric conversion is applied. Re-import if the original characters have already been lost. Formatting 124 with leading zeros later does not prove the original value was 00124.
Similarly, a date such as 04/05/2026 is ambiguous across locales. Confirm whether the source means April 5 or May 4, then use an explicit locale-aware conversion. Keep a text copy when the interpretation is unconfirmed. Compare several records against the source rather than trusting the first displayed date.
If you only have the legacy wizard
The Text Import Wizard offers Delimited mode, delimiter and text-qualifier choices, a preview, and per-column formats. In current Windows Excel it may need enabling under File > Options > Data > Show legacy data import wizards. Choose a new destination sheet so a split cannot overwrite neighboring data. Menu availability differs by edition.
Verify before cleaning or sharing
Check the header count, several quoted addresses, blank fields, the longest identifier, and a known total. Ensure the first and last expected records are present. Once the import is sound, our CSV Column Cleaner can help with supported CSV cleanup. It does not restore digits already changed by Excel or interpret an XLSX workbook. Review its preview and download a separate result so the source export remains available.
Keep these in mind
- Use the preview delimiter instead of double-clicking repeatedly.
- Keep quoted commas inside their fields.
- Protect identifiers and ambiguous dates before loading.
Common questions
Should I change Windows regional settings?
For one export, explicit import settings are usually easier to control. System-wide separator changes can affect other files and applications.
Can I get leading zeros back after saving the wrong import?
Re-import the unchanged source as Text. If the source has been overwritten and its original length is unknown, adding zeros would be guessing.
Sources & further help
Prepared with AI assistance from the linked documentation. This is a source-led explanation, not a claim of hands-on testing across every device. Read the editorial policy.