SimpleExcelExporter 1.5.0

SimpleExcelExporter

Small focused C# library for exporting .NET objects to .xlsx files. Built on top of DocumentFormat.OpenXml, it trades the SDK's full flexibility for a much simpler API aimed at one specific use case: turning a list of annotated POCOs into a spreadsheet.

The generated files open in:

• Microsoft Excel (Windows + macOS)
• LibreOffice Calc
• Google Sheets (up to its 10 M-cell quota)
• Apple Numbers — since 1.5.0 the output is strict ECMA-376 compliant

Installation

dotnet add package SimpleExcelExporter

Target framework: .NET 8.

Two ways to use it

1. Annotated POCOs (recommended for most use cases)

Annotate your model classes with attributes from SimpleExcelExporter.Annotations, then let the exporter discover the structure via reflection.

using SimpleExcelExporter;
using SimpleExcelExporter.Annotations;
using SimpleExcelExporter.Definitions;

public class Player
{
    [CellDefinition(CellDataType.String)]
    [Header(typeof(TeamRes), "PlayerNameColumnName")]
    [Index(1)]
    public string? PlayerName { get; set; }

    [CellDefinition(CellDataType.Date)]
    [Header(typeof(TeamRes), "DateOfBirthColumnName")]
    [Index(2)]
    public DateTime? DateOfBirth { get; set; }

    [CellDefinition(CellDataType.Number)]
    [Header(typeof(TeamRes), "NumberOfVictoryColumnName")]
    [Index(3)]
    public int? NumberOfVictory { get; set; }
}

public class Team
{
    private ICollection<Player>? _players;

    [SheetName(typeof(TeamRes), "SheetName")]
    [EmptyResultMessage(typeof(TeamRes), "EmptyResultMessage")]
    public ICollection<Player> Players => _players ??= new HashSet<Player>();
}

// …and then:
var team = new Team { Players = { /* … */ } };
using var stream = new FileStream("team.xlsx", FileMode.Create);
new SpreadsheetWriter(stream, team).Write();

Header labels come from a resource file (TeamRes.resx) so the same workbook can be exported in several languages.

See the full worked example at SimpleExcelExporterExample.

2. WorkbookDfn — explicit, attribute-free

If annotating your domain classes is not an option, build a WorkbookDfn by hand and pass that to the writer. You keep full control of sheet names, column ordering, cell types, and data.

var workbook = new WorkbookDfn();
var sheet = new WorksheetDfn("Players");
workbook.Worksheets.Add(sheet);

sheet.ColumnHeadings.Cells.Add(new CellDfn("Name"));
sheet.ColumnHeadings.Cells.Add(new CellDfn("Date of birth"));

var row = new RowDfn();
row.Cells.Add(new CellDfn("Alexandre", cellDataType: CellDataType.String));
row.Cells.Add(new CellDfn(new DateTime(1974, 2, 1), cellDataType: CellDataType.Date));
sheet.Rows.Add(row);

using var stream = new FileStream("players.xlsx", FileMode.Create);
new SpreadsheetWriter(stream, workbook).Write();

A more complete example is in src/ConsoleApp/Program.cs.

Supported attributes

What's in the generated XLSX?

Output conforms to ECMA-376 strict mode since 1.5.0:

• Every <c> carries r="A1"-style references (Apple Numbers requires it).
• Non-empty string cells route through xl/sharedStrings.xml (deduplicated).
• Empty cells are omitted from the sheet XML — readers infer the position (matches Excel's native output).
• Default style attributes (s="0", t="n") are omitted to keep files compact.
• docProps/core.xml is populated with dc:creator, dcterms:created, dcterms:modified.
• Deterministic writes when running under CI (ContinuousIntegrationBuild=true), plus a companion .snupkg with Source Link for step-into-source debugging.

Performance and file-size characteristics

On a 1 M-row annotated fixture (see BENCHMARK_RESULTS.md for the full table and methodology):

• Runtime is dominated by the reflection walk of annotated objects (~60 % of total) rather than the XML write.
• Files sit ~100 % larger than the pre-1.5.0 output — the price of the mandatory r="A1" attribute on every cell. Three spec-conformant optimisations (skip empty cells, omit default s="0", omit default t="n") partially offset this and keep a 1 M × 20 fixture under Google Sheets' 100 MB upload limit.

A reproducible benchmark harness is at scripts/benchmark.sh.

Repository layout

src/
  SimpleExcelExporter/         the library (published as the NuGet package)
  ConsoleApp/                  runnable harness used for manual + perf checks
test/
  SimpleExcelExporterTests/    48 NUnit tests (unit + OOXML compliance + behaviour)
scripts/
  benchmark.sh                 perf & file-size benchmark across library versions
.github/
  workflows/                   CI (on push/PR) and Release (on tag)
  dependabot.yml               weekly NuGet + Actions dependency scans

Releasing

Version bumps, NuGet publication, and GitHub Releases are driven entirely by pushing a v* tag that matches <Version> in the csproj. See RELEASING.md for the step-by-step flow, SemVer guidance, and troubleshooting.

License

LGPL-3.0-or-later.