SheetStream.NET

A zero-dependency, MIT-licensed .NET library for creating and reading Excel (.xlsx) files. Rows stream straight into the file, so 10 million cells write in seconds with flat memory.

dotnet add package SheetStream.NET
dotnet add package SheetStream.AspNetCore   # optional: one-line Excel downloads

Five-minute tour

using SheetStream;
using SheetStream.Styles;

using var package = ExcelPackage.Create("orders.xlsx");
var workbook = package.Workbook;
var header = workbook.Styles.Create(o => { o.Font.Bold = true; o.Fill.BackgroundColor = "#0A5FAD"; o.Font.Color = "#FFFFFF"; });

var sheet = workbook.Worksheets.Add("Orders");
sheet.AutoFitColumns();
sheet.FreezePanes(rows: 1);
sheet.LoadFromCollection(orders, new LoadOptions { HeaderStyle = header, TableName = "Orders" });

var chart = sheet.AddChart(ChartType.Column, "H2:P20");
chart.AddSeries(values: "C2:C13", categories: "A2:A13", name: "=C1");

package.Save();

Formulas and cached results

row.WriteFormula("SUM(C2:C101)") writes the formula and asks Excel to calculate on open. Viewers that do not calculate (file previews, Quick Look, many mobile and web viewers, ExcelReader) show such a cell blank unless the file also carries the result. Turn on sheet.CacheFormulaResults() before the first row and SheetStream computes the result itself when the formula is simple enough: numbers, references to cells already written on the same sheet, ranges, + - * /, %, parentheses and SUM, AVERAGE, COUNT, COUNTA, MIN, MAX, ROUND, ABS. Anything else (other functions, other sheets, rows not yet written, a division by zero) is written exactly as before, without a value; a wrong value is never written. The sheet keeps about 8 bytes per numeric cell to do this, which is why it is opt-in. When you already know the result, WriteFormula(formula, cachedValue) stores it directly.

Where to go next

Design in one paragraph

SheetStream writes each worksheet through an XmlWriter directly into the zip archive, so it never builds the workbook in memory. That buys flat memory and speed, and costs random access: rows are written in ascending order, one sheet at a time, and layout that Excel stores before the data (column widths, frozen panes) is declared first. Sheets that need random access can use sheet.Cells[row, column], which buffers that one sheet. Everything the library writes is validated against the ECMA-376 schema in CI and opened in real Excel before each release.

Edit this page