XLS vs CSV: Keep the Workbook, Publish the Reviewed Table

If you have an .xls workbook and need to publish its finished data, keep the workbook as your working source and make a reviewed, data-only copy for release. CSV is a useful choice when readers need one table of values rather than the workbook’s formulas, tabs and layout.

The distinction is specific: XLS is an Excel binary workbook format; CSV is a plain-text format for records and fields. Opening a CSV in Excel does not turn it into an XLS file. Microsoft explains that the original text-file format remains unchanged when opened.

XLS vs CSV at a glance

The Library of Congress describes XLS’s binary workbook structure, while RFC 4180 documents CSV’s record-and-field structure. Microsoft documents both the XLS grid limits and compatibility losses and what CSV exports retain. If Excel is displaying formulas instead of their results, it exports those formulas as text—not as a preserved workbook calculation model.

XLS is not XLSX. XLS was Excel’s default format through Excel 2003; XLSX replaced it starting with Excel 2007, as the Library of Congress source explains. Keep XLS when a recipient or legacy system specifically requires it, rather than treating it as the default choice for a new workbook. For modern workbook choices, see CSV vs XLSX.

Choose by what the recipient needs

Share a workbook when the recipient needs to inspect calculations or work across sheets. CSV does not preserve the workbook’s executable formulas or cross-sheet references. If the workbook uses newer Excel features, do not downgrade it to XLS without reviewing compatibility warnings.

The XLS row limit includes the header row. Saving a larger workbook as XLS can discard data outside its grid; converting that XLS to CSV afterward cannot recover the discarded records. Return to the complete source before preparing the release.

Share CSV when the recipient needs a finished table. It makes the release independent of spreadsheet layout: headers describe fields, and each record represents an observation. Document units, identifiers and missing-value rules separately because CSV does not declare column data types or their meaning.

Do not choose solely on file size. The useful question is whether the recipient needs the workbook or the data. Likewise, accepting an XLS upload does not establish that a publishing service will preserve its formulas, charts or formatting.

Prepare an XLS table for CSV publication

Suppose your workbook contains a calculations sheet, a notes sheet and this finished table. The following is illustrative data, not a real dataset:

site_id,report_date,site_name,visits
00123,2026-09-30,"North, Central",42
00456,2026-09-30,South,0

The intended result is two data records and four fields: text identifiers, calendar dates, site names and numeric visit counts. The comma inside North, Central belongs to the name, so the field is quoted, following CSV quoting conventions.

Use this release workflow:

1. Preserve the original workbook. Create a separate release copy containing only approved public data. Remove sensitive information; do not rely on hidden rows or sheets as a publication boundary.
2. Make one rectangular table. Use one header row, one observation per row and one field per column. Remove title blocks, merged headers, footnotes and subtotal rows from the data area. Put definitions and provenance in accompanying documentation.
3. Resolve calculations and representations. Check formula results before making a values-only release. Preserve site_id as text during entry or import, and compare it with the original identifiers: changing the format later will not restore zeros already lost. Use unambiguous dates such as 2026-09-30, keep a confirmed zero distinct from a missing value, and check that exported numbers retain the precision required for the release. Excel’s CSV export saves displayed text and values, not the workbook’s full stored representation.
4. Export the intended sheet. In Excel, use Save As and select a CSV file type. Other worksheets need separate exports: Excel saves only the current worksheet to CSV. Prefer UTF-8 when supported, and check accented or non-Latin text.
5. Inspect, then re-import. Check the raw text for the actual delimiter, intact identifiers and quoted fields. Import with explicit field types where possible, then compare headers, parsed data-record counts, sampled values and relevant totals against the release copy.

Double-clicking the exported file is not a sufficient check: Excel may reinterpret dates and leading-zero identifiers using its default import settings. Follow the CSV leading-zero checklist when exact identifiers matter.

Publish the reviewed data, not the working workbook

As of October 14, 2026, TablePage lists CSV, TSV, XLSX and XLS uploads and generates a public dataset page with a filterable table. You do not need to convert an XLS file merely because of its extension; convert when a single, inspectable table is the better release artifact.

Upload only data cleared for public release. Before distributing the page link, check the published headers, data-record count, dates, identifiers, numeric behavior and quoted text against your reviewed file. Keep the source workbook and release documentation so readers can understand where the published values came from.