Split Excel Files

1. 1

   Add your workbook

   Drag in one XLSX, XLSM, XLS, XLSB or ODS file. It is read on your device — nothing is uploaded. The sheet list, the row count of each sheet and a preview of the first rows appear straight away.

2. 2

   Choose how to split it

   By sheet turns each worksheet into its own file. By rows cuts one sheet into fixed-size batches. By column value makes one file per region, customer or category. By date makes one file per year, month or day.

3. 3

   Check the preview

   The panel names every part it is about to create and how many rows each will hold, before anything is written. Adjust the sheet, the batch size, the column or the header option until the list is what you want.

4. 4

   Split and download

   Press Split. Each part can be downloaded on its own, or take the whole set as a single ZIP. Column widths, number formats, merged cells and conditional formatting come through in every part.

Break XLSX and XLS into smaller files by sheet or row count with preview controls and ZIP export — and, now, by the value in a column or by date range as well. Split an Excel workbook by sheet or row count into separate files, choose which sheets to include, watch a live preview of the sheet you are cutting, and see the exact list of files the run will produce before a single one is written. Everything happens on your device: the workbook is never uploaded, there is no queue to wait in, and no copy is left on anyone's server afterwards.

• Four ways to split. One file per sheet, fixed batches of rows, one file per distinct value in a column, or one file per year, month or day of a date column. All four are available on the free plan.
• A sheet picker, not all-or-nothing. Tick the sheets you want. A twelve-tab workbook can produce the three files you need instead of twelve you have to sift through.
• A real preview. The first rows of the chosen sheet are shown as a grid, with the splitting column highlighted, and the panel below names every part the run will create and how many rows each will hold.
• Formatting that survives. Column widths, number and currency formats, fonts, fills, merged cells, conditional formatting and data validation are carried into every part.
• Results you can inspect. Each part is listed with its sheet, its row count and its size, downloadable on its own or as one ZIP for the whole set.
• Names that behave. Sheet names and cell values become filenames safely — separators, control characters and reserved Windows names are neutralised, and two parts can never quietly overwrite each other.

This is the question that decides whether a splitter is useful, and most tools answer it vaguely. An XLSX file is not really a spreadsheet — it is a small archive of XML documents, and almost everything that makes a sheet look the way it does lives outside the cells themselves. Column widths sit in one element, the number and currency formats in a shared styles document, the conditional formatting rules in another, and each cell simply points at a style by number. A tool that reads a workbook into memory as values and writes a new one back out therefore has to reconstruct all of that, and most cannot, which is why parts so often arrive in a default font with dates showing as five-digit numbers.

• Split by sheet is exact. The worksheet is copied across without ever being decoded, along with the shared styles, the theme, the shared text table and any images. Charts, pictures, hyperlinks, autofilters, frozen panes, print areas and formulas all arrive unchanged, because nothing about them was ever interpreted.
• Row, column and date splits keep the appearance. Each row is moved with its style references intact, so number formats, fonts and fills carry over. Column widths come with the sheet. Merged cells, conditional formatting and data validation ranges are re-addressed to the new row numbers rather than dropped.
• Sliced parts drop what cannot be re-addressed. Autofilters, drawings, embedded tables and sheet-level hyperlinks refer to positions in the original sheet, so a slice leaves them out rather than carrying a reference that points nowhere. Use split by sheet when those matter — there they are kept exactly.
• Merges that a slice would break are removed, not resized. A merged block spanning rows 3 to 5 has no honest equivalent in a part that kept rows 3 and 91, so it is dropped. One that survives intact moves with its rows.
• XLS, XLSB and ODS carry values, not appearance. None of the three is an XML package, so each is converted to XLSX first. Every value, every number format and the sheet structure survive; the original styling does not. The badge next to the filename says which path your file took before you run.

A 24.7 MB workbook of six sheets and 60,000 rows is read in 70 ms and split by sheet in 46 ms, on your device.

One file per worksheet is the split most people arrive for: a workbook has grown a tab per department, per month or per client, and each one now needs to go to a different person. Tick the sheets you want and each becomes a standalone workbook named after the sheet it came from.

The result is a genuine single-sheet workbook rather than a copy of the original with the other tabs hidden. The sheet list, the relationships between the parts of the file and the content-type declarations are all rebuilt for one sheet; workbook-level names that referred to sheets you did not keep are removed so nothing resolves to a reference error; and the recalculation cache — a workbook-wide index of every formula cell — is dropped, with Excel told to rebuild it on open. That last detail is why the parts open cleanly instead of showing a repair prompt, and it is the step most hand-rolled splitters miss.

Fixed batches are what an import limit wants. Systems that accept spreadsheet uploads often cap them — 5,000 rows, 10,000 rows, 65,536 rows for anything still speaking the old format — and a 48,000-row export has to be cut into pieces that each fit.

Set the batch size and choose whether the header row is repeated at the top of every part; almost always it should be, because an importer that reads column names from the first row will otherwise treat the first data row of every part after the first as headings. The batch-size control's minimum moves with the sheet and with your plan, so it can never be set to a value that would produce more files than the run is allowed to write. That sounds like a small thing and is not: a 200,000-row sheet asked for one row per file is 200,000 files, and a tool that offers the setting is a tool that will one day be asked for it.

One file per region, per customer, per category, per branch. This is the split people otherwise do by hand — filter the column, copy the visible rows, paste into a new workbook, repeat — and it is both the slowest and the easiest to get subtly wrong.

Choose Column value and pick the column from the list, which names each one by its header text rather than only by its letter, so you are choosing "region" and not "column C". Every distinct value in that column becomes its own file, named after the value, holding exactly the rows that carried it, with the header repeated if you asked for it. Rows whose cell is empty are collected into their own labelled part rather than silently discarded — a blank is a group, and losing those rows is how a split quietly loses data. The parts come out in the order the values first appear in the sheet, so a column that was already sorted produces files in the order you expect.

A log, an export or a ledger with a date column usually wants to be cut by period rather than by size: one file per year for archiving, one per month for reporting, one per day for reconciliation.

Choose Date range, pick the date column, and choose year, month or day. The one thing that makes this reliable is how a date is recognised at all. A spreadsheet has no date type — a date is a plain number counting days, and the only reason it appears as a date on screen is that its cell's number format says so. This tool reads that format, which is why a column of dates is grouped as dates and a column of quantities that happen to look similar is not, and why nothing ever comes out of a split showing 45658 where a date should be. That last failure is exactly what the previous version of this page produced, because it read the values and ignored the formats.

Every split has to answer this, and "we keep your formulas" is usually the wrong answer rather than the generous one.

Split by sheet keeps formulas exactly as they were, because the whole sheet travels intact — a formula that referred to a cell on the same sheet still refers to it. The row, column and date modes replace each formula with the value it was last showing. That is not a compromise; it is the only correct result. A formula such as =SUM(A2:A5000) placed into a file that contains only rows 1 to 1000 no longer describes the same sum, and a reference to a row the part does not contain is an error message where a number used to be. Excel stores each formula's most recent result immediately beside the formula, so the number written into the part is precisely what the sheet was displaying — no re-calculation, no approximation, no dependence on how the file is later opened. Where a formula produced text rather than a number, the text is written as a proper text cell so the part opens without complaint.

Five formats go in and every part comes out as XLSX, or as CSV on the Pro plan. Two of the five take an exact path; the other three are converted first, and the difference is stated on screen the moment the file is read rather than discovered afterwards.

XLSX and XLSM — the exact path

These are XML packages, so the tool works on the file's own parts. Styles, themes, shared text and images are copied across as they are, and split by sheet reproduces the worksheet byte for byte. Macros in an XLSM are not carried into the parts, because a macro belongs to the workbook rather than to a sheet; everything else is.

XLS, XLSB and ODS — converted first

A 97–2003 XLS is a binary record stream, an XLSB is a binary package, and an ODS is a different XML vocabulary entirely. None can be spliced, so each is read and rewritten as XLSX before the split. Values, types and number formats survive; the original fonts, fills and widths do not. The badge beside the filename reads "values only" for these three so the trade-off is visible before you run rather than after.

What is refused, and where to go instead

A password-protected workbook is refused with a message that says so, rather than opening blank — remove the password in Excel first. A file that is not a workbook at all, including a CSV that has been renamed, is refused with a link to the converter that does take it. An empty file is refused as empty. In every case the refusal names what the file was, what the tool takes, and one thing you can do about it; you should never reach a progress bar that then fails.

All four split modes, the sheet picker, the preview, the header option and the ZIP download are free. The free plan caps scale, not capability.

• Free. One workbook at a time, up to 25 MB, producing up to 50 files per run, written as XLSX and named after the workbook and the sheet.
• Pro. Up to 20 workbooks queued, sized by what your device can handle rather than by a number we picked, producing up to 2,000 files per run, with CSV as an output option, a custom filename pattern and a custom name for the ZIP.
• Both. Every split mode, every preview, the whole result list and the single-ZIP download. Nothing that decides whether the tool works for you is behind the paywall.

• Merge Excel
• Excel Viewer
• Excel to CSV
• Excel to JSON
• Excel to PDF
• CSV to Excel