ExcelMapper

ExcelMapper

A library to map POCO objects to Excel files.

Features

• Read and write Excel files
• Uses the pure managed NPOI library instead of the Jet database engine
• Map to Excel files using header rows (column names) or column indexes (no header row)
• Map nested objects (parent/child objects)
• Optionally skip blank lines when reading
• Preserve formatting when saving back files
• Optionally let the mapper track objects
• Map columns to properties through convention, attributes or method calls
• Use custom or builtin data formats for numeric and DateTime columns
• Map formulas or formula results depending on property type
• Map JSON
• Fetch/Save dynamic objects
• Use records
• Provide custom object factories

Read objects from an Excel file

var products = new ExcelMapper("products.xlsx").Fetch<Product>();

This expects the Excel file to contain a header row with the column names. Objects are read from the first worksheet. If the column names equal the property names (ignoring case) no other configuration is necessary. The format of the Excel file (xlsx or xls) is autodetected.

Map to specific column names

public class Product
{
  public string Name { get; set; }
  [Column("Number")]
  public int NumberInStock { get; set; }
  public decimal Price { get; set; }
}

This maps the column named Number to the NumberInStock property.

Map to column indexes

Column indexes start at 1.

public class Product
{
    [Column(1)]
    public string Name { get; set; }
    [Column(Letter="C")]
    public int NumberInStock { get; set; }
    [Column(4)]
    public decimal Price { get; set; }
}

var products = new ExcelMapper("products.xlsx") { HeaderRow = false }.Fetch<Product>();

Note that column indexes don’t need to be consecutive. When mapping to column indexes, every property needs to be explicitly mapped through the ColumnAttribute attribute or the AddMapping() method. You can combine column indexes with column names to specify an explicit column order while still using a header row.

Map through method calls

var excel = new ExcelMapper("products.xls");
excel.AddMapping<Product>("Number", p => p.NumberInStock);
excel.AddMapping<Product>(1, p => p.NumberInStock);
excel.AddMapping(typeof(Product), "Number", "NumberInStock");
excel.AddMapping(typeof(Product), ExcelMapper.LetterToIndex("A"), "NumberInStock");

Multiple mappings

You can map a single column to multiple properties but you need to be aware of what should happen when mapping back from objects to Excel. To specify the single property you want to map back to Excel, add MappingDirections.ExcelToObject in the Column attribute of all other properties that map to the same column. Alternatively, you can use the FromExcelOnly() method when mapping through method calls.

public class Product
{
    public decimal Price { get; set; }
    [Column("Price", MappingDirections.ExcelToObject)]
    public string PriceString { get; set; }
}

// or

excel.AddMapping<Product>("Price", p => p.PriceString).FromExcelOnly();

Column attributes are inherited by default, resulting in multiple mappings for a single overridden property if you add a Column attribute to the property in base and derived classes. To prevent this, set the Inherit property to false on the Column attribute in the base class.

Dynamic mapping

You don’t have to specify a mapping to static types, you can also fetch a collection of dynamic objects.

var products = new ExcelMapper("products.xlsx").Fetch(); // -> IEnumerable<dynamic>
products.First().Price += 1.0;

The returned dynamic objects are instances of ExpandoObject with an extra property called __indexes__ that is a dictionary specifying the mapping from property names to column indexes. If you set the HeaderRow property to false on the ExcelMapper object, the property names of the returned dynamic objects will match the Excel “letter” column names, i.e. “A” for column 1 etc.

Save objects

var products = new List<Product>
{
    new Product { Name = "Nudossi", NumberInStock = 60, Price = 1.99m },
    new Product { Name = "Halloren", NumberInStock = 33, Price = 2.99m },
    new Product { Name = "Filinchen", NumberInStock = 100, Price = 0.99m },
};

new ExcelMapper().Save("products.xlsx", products, "Products");

This saves to the worksheet named “Products”. If you save objects after having previously read from an Excel file using the same instance of ExcelMapper the style of the workbook is preserved allowing use cases where an Excel template is filled with computed data.

Track objects

var products = new ExcelMapper("products.xlsx").Fetch<Product>().ToList();
products[1].Price += 1.0m;
excel.Save("products.out.xlsx");

Ignore properties

public class Product
{
    public string Name { get; set; }
    [Ignore]
    public int Number { get; set; }
    public decimal Price { get; set; }
}

// or

var excel = new ExcelMapper("products.xlsx");
excel.Ignore<Product>(p => p.Price);

Use specific data formats

public class Product
{
    [DataFormat(0xf)]
    public DateTime Date { get; set; }

    [DataFormat("0%")]
    public decimal Number { get; set; }
}

You can use both builtin formats and custom formats. The default format for DateTime cells is 0x16 (“m/d/yy h:mm”) and can be changed through the properties DefaultDateFormat and DateFormat.

Explicitly set data formats override existing formats in an Excel file. Otherwise existing column and cell styles will be left untouched.

Map formulas or results

Formula columns are mapped according to the type of the property they are mapped to: for string properties, the formula itself (e.g. “A1+B1”) is mapped, for other property types the formula result is mapped. If you need the formula result in a string property, use the FormulaResult attribute.

public class Product
{
    [FormulaResult]
    public string Result { get; set; }
}

// or

excel.AddMapping<Product>("Result" p => p.Result).AsFormulaResult();

If you want to save formulas you need to use the FormulaAttribute attribute or call AsFormula() if mapping manually. It’s not needed if you only want to map from Excel to objects (deserialize).

public class Product
{
    [Formula]
    public string Formula { get; set; }
}

// or

excel.AddMapping<Product>("Formula" p => p.Formula).AsFormula();

☝️ The string values of formula properties must not start with the = sign. So instead of =A1+B1 set the property’s value to A1+B1.

Custom mapping

If you have specific requirements for mapping between cells and objects, you can use custom conversion methods. Here, cells that contain the string “NULL” are mapped to null:

public class Product
{
    public DateTime? Date { get; set; }
}

excel.AddMapping<Product>("Date", p => p.Date)
    .SetCellUsing((c, o) =>
    {
        if (o == null) c.SetCellValue("NULL"); else c.SetCellValue((DateTime)o);
    })
    .SetPropertyUsing(v =>
    {
        if ((v as string) == "NULL") return null;
        return Convert.ChangeType(v, typeof(DateTime), CultureInfo.InvariantCulture);
    });

Header row and data row range

You can specify the row number of the header row using the property HeaderRowNumber (default is 0). The range of rows that are considered rows that may contain data can be specified using the properties MinRowNumber (default is 0) and MaxRowNumber (default is int.MaxValue). The header row doesn’t have to fall within this range, e.g. you can have the header row in row 5 and the data in rows 10-20.

JSON

You can easily serialize to and from JSON formatted cells by specifying the Json attribute or AsJson() method.

public class ProductJson
{
    [Json]
    public Product Product { get; set; }
}

// or

var excel = new ExcelMapper("products.xls");
excel.AddMapping<ProductJson>("Product", p => p.Product).AsJson();

This also works with lists.

public class ProductJson
{
    [Json]
    public List<Product> Products { get; set; }
}

Name normalization

If the header cell values are not uniform, perhaps because they contain varying amounts of whitespace, you can specify a normalization function that will be applied to header cell values before mapping to property names. This can be done globally or for specific classes only.

excel.NormalizeUsing(n => Regex.Replace(n, "\s", ""));

This removes all whitespace so that columns with the string " First Name " map to a property named FirstName.

Records

Records are supported. If the type has no default constructor (as is the case for positional records) the constructor with the highest number of arguments is used to initialize objects. This constructor must have a parameter for each of the mapped properties with the same name as the corresponding property (ignoring case). The remanining parameters will receive the default value of their type.

Nested objects

Nested objects are supported and should work out of the box for most use cases. For example, if you have a sheet with columns Name, Street, City, Zip, Birthday, you can map to the following class hierarchy without any configuration:

public class Person
{
    public string Name { get; set; }
    public DateTime Birthday { get; set; }
    public Address Address { get; set; }
}

public class Address
{
    public string Street { get; set; }
    public string City { get; set; }
    public string Zip { get; set; }
}

var customers = new ExcelMapper("customers.xlsx").Fetch<Person>();

This works with records, too:

public record Person(string Name, DateTime Birthday, Address Address);
public record Address(string Street, string City, string Zip);

Object factories

You can specify a custom object factory for any type which will be used to create object instances for mapped properties of that type. This can be useful to handle cases where object creation is otherwise not possible (such as for properties that have interface types) or where you want to execute specific initialization logic.

public class Person
{
    public string Name { get; set; }
    public IAddress Address { get; set; }
}

excel.CreateInstance<IAddress>(() => new Address());