csv-parser/csvparser.go

package csvparser

import (
	"fmt"
	"strings"
	"unicode/utf8"
)

type CsvParser struct {
	enclosers  []string
	delimiters string
  mergeDelimiters string
	fields     []string
}

/*
* delimiters: string with all delimiter chars
* mergeDelimiters: if true, all successive delimiters are considered as one
* enclosers: array of string, each string contains a enclorser tuple: "", [], {}...
* lineFormat: format of the line (each fieldname or ignore)
*/
func (p *CsvParser) Initialize(delimiters string, mergeDelimiters bool, enclosers []string, lineFormat string) error {
	if utf8.RuneCountInString(delimiters) == 0 {
		return fmt.Errorf("delimiter shoud be at least one character")
	}
	p.delimiters = delimiters
  p.mergeDelimiters = mergeDelimiters
	for _, encloser := range enclosers {
		if utf8.RuneCountInString(encloser) != 2 {
			return fmt.Errorf("encolser should have to characters")
		}
	}
	p.enclosers = enclosers
	// line format is in the form of: field1 field2 ignore ...
	// if field name is ignore, it is parsed but not retained
	p.fields = strings.Split(lineFormat, " ")

	return nil
}

type parserState struct {
  delimiter bool
  enclosed bool
  encloserEnd rune
}

func (p *CsvParser) Parse(line string) (map[string]string, error) {
	currentFieldIndex := 0
	valueStart := -1
	escape := false
	enclosed := false

	ret := make(map[string]string)
	indexMax := len(line) - 1
	maxFieldIndex := len(p.fields) - 1
	delimiters := p.delimiters
  delimiter := false
	for index, r := range line {
		if r == '\\' {
			// Check if EOL before continue
			escape = true
			continue
		}
		if escape {
			escape = false
			continue
		}
		for _, d := range delimiters {
      if r == d {
        delimiter = true
        break
      }
		}
    if delimiter && valueStart > -1 {
			if p.fields[currentFieldIndex] != "ignore" {
				ret[p.fields[currentFieldIndex]] = line[valueStart:index]
			}
			currentFieldIndex++
    }
		if delimiter {
			if p.fields[currentFieldIndex] != "ignore" {
				ret[p.fields[currentFieldIndex]] = line[valueStart:index]
			}
			currentFieldIndex++
			if currentFieldIndex > maxFieldIndex {
				break
			}
			valueStart = index + 1
			if enclosed {
				enclosed = false
				delimiters = p.delimiters
				// Omit next delimiter
				escape = true
			}
			continue
		}

		if index >= indexMax {
			if p.fields[currentFieldIndex] != "ignore" {
				ret[p.fields[currentFieldIndex]] = line[valueStart:]
			}
			continue
		}

		for _, encloser := range p.enclosers {
			runes := []rune(encloser)
			if r == runes[0] {
				// opening encloser
				enclosed = true
				delimiters = string(runes[1])
				valueStart++
				break
			}
		}
	}
	return ret, nil
}