A customer code starts as 000742 in an old Excel workbook. After a CSV export, someone opens it and sees 742. The file still loads. Every row looks tidy. Yet that tiny change can break a customer lookup, reject an import, or attach records to the wrong account. A reliable export needs more than a successful save. You need to check what Excel stores, what the CSV contains, and what the receiving application reads.
Keep the Characters That Count
Treat identifiers as text, give dates an agreed format, and confirm the delimiter before exporting. Check the CSV in a text editor before opening it in another spreadsheet. Leading zeros can disappear during preparation, conversion, or import, and each stage needs its own check. Keep an untouched workbook, test a small sample, and compare the finished file with known records before sharing it. CSV carries characters, but it does not carry Excel’s formatting rules.
Check What the Workbook Actually Stores
Legacy workbooks often mix real values with display formatting. A cell showing 000742 might contain the number 742 with a custom format of 000000. Another cell might contain the six text characters 000742. Those cells look alike, but an export tool may handle them differently.
Select a few identifier cells and inspect their formulas, values, and number formats. Check cells from different parts of the worksheet. Old files may have years of manual edits, pasted records, and formulas mixed into one column. One correct example does not prove that the whole column is consistent.
Keep the original workbook untouched and prepare an export copy. Confirm which worksheet contains the records you need. CSV cannot hold multiple worksheets, cell styles, or workbook formulas as working spreadsheet features. A normal CSV export writes field values from a selected sheet.
Preserve Identifiers Before You Export
Postal codes, account references, product codes, and telephone numbers may contain digits without representing quantities. You do not add two customer identifiers together. Their exact characters matter, including any zeros at the beginning.
Turn Display Padding Into Actual Text
If a six-character identifier is stored as a number, a helper column can produce the text you need. For a value in cell A2, =TEXT(A2,"000000") produces 000742 from 742. Fill the formula down, check the results, then copy the helper column and paste its values into a destination column already formatted as Text.
Use this method only when six characters is a confirmed rule. Adding zeros to every code can damage a column containing valid identifiers of different lengths. A missing zero cannot be reconstructed reliably unless you know the required format or have an authoritative original record.
Changing an existing numeric cell’s format to Text does not restore characters already lost. Prepare text columns before entering or importing identifiers, or create verified text values from the existing data.
Check Long Codes for Earlier Damage
Long numeric identifiers deserve a separate inspection. Excel supports only 15 significant digits for numeric values. If a longer code was entered as a number, later digits may already have changed. Formatting that damaged number as text will preserve the damaged version.
Compare long account numbers and transaction references against their original source. If the workbook has already altered them, recover the original text before exporting. A converter cannot infer missing digits.
Make Dates Clear Before They Leave Excel
A date written as 03/04/2008 can mean March 4 or April 3. CSV does not record which interpretation the writer intended. A receiving application may choose based on its own regional settings.
Agree on a date format with whoever will import the file. For calendar dates, YYYY-MM-DD is a clear choice when supported by the destination. In Excel, a helper formula using =TEXT(B2,"yyyy-mm-dd") can create that output from a valid date value.
Check that B2 really contains a date first. Applying a date format to an ambiguous text entry does not establish what that entry means. Inspect examples with days greater than 12 and compare them with known records. Leave missing dates empty unless the receiving system explicitly requires another value.
Very old workbooks may use different date systems. Excel’s 1900 and 1904 date systems differ by 1,462 days. If dates have shifted by roughly four years after copying between workbooks, investigate that setting before preparing the CSV.
For timestamps, settle the time zone separately. An Excel date and time value does not automatically include a time zone. Add an offset or separate zone field only when you know the original meaning and the destination accepts that representation.
Choose a Delimiter the Receiving Application Accepts
A CSV extension does not guarantee that every exporter writes commas. Regional settings and application options may produce semicolons instead. Confirm the required separator before converting the workbook, then inspect the actual output.
A file using the wrong delimiter may appear as one wide column during import. Another warning sign is a record split into too many fields. Both problems can happen even when the file opens without an error message.
Fields containing commas, quotation marks, or line breaks need careful handling. In conventional comma-separated CSV, a field containing a comma belongs inside double quotes. A quotation mark inside that field is represented by two double quotes. For example, the value Valve, 2" inlet becomes "Valve, 2"" inlet".
Do not replace every comma in the exported text to change separators. Some commas belong inside addresses, descriptions, or quoted fields. Use a converter or importer that understands field boundaries.
Double quotes protect CSV structure. They do not reliably tell a spreadsheet to treat an identifier as text. A spreadsheet can still interpret "000742" as a number during import unless you control the column type.
Test the Export With Representative Records
A small trial helps you catch problems before processing an entire archive. Include records with leading zeros, long identifiers, blank cells, accented names, commas, quotation marks, and dates with known meanings. Include a multiline note if the workbook contains them.
Record the expected values. Write down several exact identifiers and dates from verified source records. Note the expected number of data records and columns.
Prepare the export sheet. Convert verified identifier fields to text values. Standardize dates, confirm headers, and remove content that does not belong in the transfer.
Export the sample. Select the intended worksheet, separator, and character encoding. Save to a new filename so the original workbook remains available.
Inspect the raw file. Open the CSV in a text editor. Search for the expected leading zeros and check dates, separators, quotes, and accented characters.
Import with explicit column types. Load identifiers as Text in the receiving application. Confirm that dates and numeric quantities follow the agreed rules.
Compare before proceeding. Check record counts, column counts, exact identifiers, and a numeric total where useful. Repeat the same checks on the complete export.
If you need an XLS to CSV converter, apply these checks to its output before using the full dataset. Test whether it exports displayed formatting or underlying values, which sheet it selects, and how it handles formulas. Verify those behaviors with sample records.
Inspect the CSV Without Triggering Another Conversion
Double-clicking a CSV often opens it in a spreadsheet that immediately guesses column types. That can remove leading zeros from the view even when the exported file contains them. Saving after that conversion can then write the altered values back to disk.
Start with a text editor. Search for a complete identifier, including its zeros. If the raw file contains 000742 but a spreadsheet shows 742, investigate the import settings. If the raw file already contains 742, investigate the source preparation or export process.
When loading the file into Excel, use an import workflow that lets you assign identifier columns as Text before loading. Available controls vary by version. In workflows that infer types automatically, review the conversion steps before accepting the result.
Avoid adding spreadsheet formulas to CSV fields just to force a particular display. That changes the exported content and may create compatibility problems. Explicit import types provide a cleaner way to preserve identifiers.
Check Encoding, Record Counts, and Private Fields
Encoding controls how characters become file bytes. UTF-8 is a useful choice when the destination supports it, but confirm the requirement for older systems. Check names and addresses containing accented letters or non-Latin characters. A file that preserves every zero can still damage a customer’s name.
Count parsed records rather than physical text lines. A quoted field containing a line break can span several lines while remaining part of one record. Check that every parsed record has the expected number of fields, including records with empty values at the end.
Compare useful totals as an additional check. A quantity total or invoice total can reveal missing records, but it cannot prove that identifiers survived. Use totals alongside exact text comparisons.
Before uploading a legacy workbook to any online service, review what it contains and whether uploading is allowed. Work from a reduced copy containing only the required data. Check hidden sheets, unused columns, and personal information. For restricted records, use an approved local export process.
Send a File Whose Meaning Survives
A dependable CSV handoff includes a short description of the delimiter, encoding, date format, and columns that must be imported as text. State whether there is a header row and how blank values should be handled.
Keep the verified CSV alongside the original workbook and your export notes. Before sharing it, check a leading-zero identifier, a long code, a known date, a field containing punctuation, and a name containing accented characters. Those small checks help confirm that the exported file carries the records you intended to send.