JoshClose
commented
2 years ago There are already [Name(string name)] and [Name(params string[])] attributes to do this. You should still supply and Index because the order of the properties returned from Type.GetProperties() may not be the same as the order they appear in your class. If you have a header you may not care about the order.
I'm not going to add any features around multiple headers because it's not a valid CSV file. It's pretty easy to do yourself though, and you can put the functionality in an extension method or static method for reuse.
void Main()
{
var records = new List<Foo>
{
new Foo { feed_product_type = "ABIS_BOOK" },
};
var config = new CsvConfiguration(CultureInfo.InvariantCulture)
{
};
using (var writer = new StringWriter())
using (var csv = new CsvWriter(writer, config))
{
csv.WriteHeader<Foo>();
csv.NextRecord();
foreach (var property in typeof(Foo).GetProperties())
{
csv.WriteField(property.Name);
}
csv.NextRecord();
csv.WriteRecords(records);
writer.Dump();
}
}
private class Foo
{
[Name("Product Type")]
public string feed_product_type { get; set; }
}
booktrakker
I came to realize after you answered my question about how to handle multiple headers has an elegant solution I think would be a real boon. I suspect this would be rather easy to implement so I do hope you find it worth implementing.
Basically you would add support for a new Attribute - something like [Header("headertext")] - that we would add to each property in much the same way you had me use an index. This Attribute would put the required header in the row before the header that is generated by the property name, and if done right, would allow multiple such attributes to be placed before each property to account for however many rows are needed.
An initial pass would get the max number of such rows so that any property with fewer rows can be populated with blank entries, all other columns would have the Attributes in the rows before the Property Name header row.
This would make it extremely easy to maintain without having to index the rows, which can become a maintenance issue.
In my Amazon example the class that defines the structure would have an entry something like this:
[Header("Product Type")] public string feed_product_type { get; set; }
This would result in the Column Header looking like this (where "ABIS_BOOK" is the actual value of that field for the first record exported):
Product Type feed_product_type "ABIS_BOOK"