gocsv

By default, gocsv prefixes nested struct fields with the parent field's name (e.g., Parent.ChildField). To inline a nested struct and use its fields directly as top-level columns without a prefix, use the csv:"." tag on the nested field.

type Foo struct {
	B Bar `csv:"."` // Fields in Bar will be 'a, b' instead of 'B.a, B.b'
	X int `csv:"x"` 
}

type Bar struct {
	A int `csv:"a"` 
	B int `csv:"b"` 
}

You can define custom serialization and deserialization logic for specific types by implementing the MarshalCSV() (string, error) and UnmarshalCSV(csv string) error methods on your type. This allows you to handle custom date formats or complex data transformations during the CSV process.

type DateTime struct {
	time.Time
}

// Convert the internal date as CSV string
func (date *DateTime) MarshalCSV() (string, error) {
	return date.Time.Format("20060201"), nil
}

// Convert the CSV string as internal date
func (date *DateTime) UnmarshalCSV(csv string) (err error) {
	date.Time, err = time.Parse("20060201", csv)
	return err
}

Install the gocsv package using go get:

go get -u github.com/gocarina/gocsv

To convert a slice of structs into a CSV string, use gocsv.MarshalString. To save the data directly to a file, use gocsv.MarshalFile.

// Get all clients as CSV string
csvContent, err := gocsv.MarshalString(&clients)

// Use this to save the CSV back to the file
err = gocsv.MarshalFile(&clients, clientsFile)

Use gocsv.UnmarshalFile to load CSV data from an os.File into a slice of pointers to a struct. Use struct tags with the csv key to map CSV column headers to struct fields. To ignore a field, use the csv:"-" tag.

type Client struct {
	Id            string `csv:"client_id"` 
	Name          string `csv:"client_name"` 
	Age           string `csv:"client_age"` 
	NotUsedString string `csv:"-"` 
	NotUsedStruct NotUsed `csv:"-"` 
}

// ...

clients := []*Client{}
if err := gocsv.UnmarshalFile(clientsFile, &clients); err != nil {
	panic(err)
}

You can override the default CSV reader and writer behavior using gocsv.SetCSVReader and gocsv.SetCSVWriter. This is useful for changing delimiters (e.g., using a pipe | or dot .), enabling LazyQuotes, or using gocsv.LazyCSVReader.

// Configure a custom reader with a pipe delimiter
gocsv.SetCSVReader(func(in io.Reader) gocsv.CSVReader {
    r := csv.NewReader(in)
    r.Comma = '|'
    return r
})

// Configure a custom writer with a pipe delimiter
gocsv.SetCSVWriter(func(out io.Writer) *gocsv.SafeCSVWriter {
    writer := csv.NewWriter(out)
    writer.Comma = '|'
    return gocsv.NewSafeCSVWriter(writer)
})

The SafeCSVWriter provides a thread-safe wrapper around the standard library's *csv.Writer. It uses a sync.Mutex to ensure that concurrent calls to Write and Flush do not cause race conditions or data corruption when multiple goroutines attempt to write to the same CSV destination simultaneously.

To use it, wrap an existing *csv.Writer using NewSafeCSVWriter(original).

You can override the default CSV reader and writer used by the Marshal and Unmarshal functions.

Customizing the Writer: Use SetCSVWriter(csvWriter func(io.Writer) *SafeCSVWriter) to provide your own writer factory. The default is DefaultCSVWriter, which uses csv.NewWriter wrapped in a SafeCSVWriter.

Customizing the Reader: Use SetCSVReader(csvReader func(io.Reader) CSVReader) to provide your own reader factory.

• DefaultCSVReader(in io.Reader): Returns a standard csv.NewReader.
• LazyCSVReader(in io.Reader): Returns a reader with LazyQuotes and TrimLeadingSpace enabled.