vchezganov / gocsv

Simple CSV unmarshaller for Go structs

gocsv

Go package for parsing CSV records into structs. Currently, it supports only the following types and its references:

• int, int8, int16, int32, int64
• uint, uint8, uint16, uint32, uint64
• float32, float64
• string

In addition, you may provide own function to be used for parsing values. The function should accept string parameter and return error if there are any errors when parsing.

Example

package main

import (
	"fmt"
	"io"
	"strings"

	"github.com/vchezganov/gocsv"
)

type Person struct {
	Name string `csv:"name"`
	Age  uint   `csv:"age"`
	ID   string `csv:"pass,ParseID"`
}

func (p *Person) ParseID(value string) error {
	p.ID = fmt.Sprintf("ABC-%s", value)
	return nil
}

func main() {
	stringReader := strings.NewReader("age,pass,name\n32,12345,Vitaly\n45,54321,Alexey")

	reader, err := gocsv.NewReader[Person](stringReader)
	if err != nil {
		panic(err)
	}

	for {
		model, err := reader.Next()
		if err == io.EOF {
			break
		} else if err != nil {
			continue
		}

		fmt.Printf("Person: %v\n", *model)
	}
}

Or Marshaller could be used directly:

package main

import (
	"encoding/csv"
	"errors"
	"fmt"
	"strconv"
	"strings"

	"github.com/vchezganov/gocsv"
)

type Person struct {
	Name     string `csv:"name"`
	Age      int    `csv:"age"`
	Location string `csv:"city"`
	ID       int    `csv:"passport,ParseID"`
}

func (p *Person) ParseID(value string) error {
	s, err := strconv.Atoi(value)
	if err != nil {
		return err
	}

	if 10000 <= s && s <= 99999 {
		p.ID = s
		return nil
	}

	return errors.New("ID is not valid")
}

func main() {
	s := strings.NewReader("name,age,city,passport\nVitaly,25,Bonn,10000")
	csvReader := csv.NewReader(s)
	headers, err := csvReader.Read()
	if err != nil {
		panic(err)
	}

	marshaller, err := gocsv.NewMarshaller(headers)
	if err != nil {
		panic(err)
	}

	model := new(Person)
	records, err := csvReader.Read()
	if err != nil {
		panic(err)
	}

	err = marshaller.Unmarshal(records, model)
	if err != nil {
		panic(err)
	}

	fmt.Printf("Person: %v\n", *model)
}

Next steps

• Support composition

type A struct {
	ID        int    `csv:"id"`
	Timestamp string `csv:"timestamp"`
}

// id, timestamp, name
type B struct {
	*A
	Name int `csv:"name"`
}

// prefix_id, prefix_timestamp, name
type C struct {
	Base *A  `csv:"prefix"`
	Name int `csv:"name"`
}

• Parsing function to accept not only string but int, float, etc.
• Converting structs into CSV dictionary, slice or string