artonge / go-csv-tag Goto Github PK

The library fails to parse CSV files that contain lines with trailing spaces. I came across this issue when trying to read a GTFS file using the go-gtfs library.

E.g.:

header1, header2, header3
line1, 1, 1.2                                                                                                                                                                                                                                                                                                                         
line2, 2, 2.3                                                                                                                                                                                                                                                                                                                         
line3, 3, 3.4           

Error produced:

error mapping csv from path csv_files/trailingSpaces.csv:
        ==> error mapping the content to the destination
        ==> line: [line1 1 1.2                ] to slice: {line1 1 0}:
        ==> error parsing float '1.2                ':
        ==> strconv.ParseFloat: parsing "1.2                ": invalid syntax

Hi I was trying to use this package to decode a struct with int64 fields and It wasn't working. There is a simple fix for this in load.go:

func storeValue(rawValue string, valRv reflect.Value) error {
	switch valRv.Kind() {
	case reflect.String:
		valRv.SetString(rawValue)
	case reflect.Int64: // Add this.
		fallthrough
	case reflect.Int:
		value, err := strconv.ParseInt(rawValue, 10, 64)
		if err != nil && rawValue != "" {
			return fmt.Errorf("error parsing int '%v':\n	==> %v", rawValue, err)

		}

It would be nice if we could specify a custom tag key. An example use case is where we might want to make use of existing tags such as json. A working example here

When Dumping structs containing floats, the default representation is used. For high exponents, this is scientific notation, e.g. 1e-06 for 0.000001.

This is incompatible with some specifications that use csv. Currently, there's no way of specifying which representation go-csv-tag should use. Similar situations may arrise for other types, like if an integer should be duped as a hexadeximal value.

First use the example DumpToFile to generate a csv test file, then use LoadFromPath to load this file, panic happened

generated csv file :
name, ID, number
'name1', 1, 1.2

panic: runtime error: index out of range [1] with length 1

goroutine 1 [running]:
github.com/artonge/go-csv-tag.mapToDestination(0xc0000c7f48, 0x3, 0x3, 0xc0000b4
330, 0x2, 0x2, 0x4bbf00, 0xc000096560, 0x3, 0xc0000b4330)
E:/Go/gopath/src/github.com/artonge/go-csv-tag/load.go:182 +0xe9f
github.com/artonge/go-csv-tag.LoadFromReader(0x503ac0, 0xc0000ca028, 0x4bbf00, 0
xc000096560, 0xc0000c7f28, 0x1, 0x1, 0xc00008c8f0, 0xc0000a00c0)
E:/Go/gopath/src/github.com/artonge/go-csv-tag/load.go:48 +0x17e
github.com/artonge/go-csv-tag.LoadFromPath(0x4e52c4, 0x11, 0x4bbf00, 0xc00009656
0, 0xc0000c7f28, 0x1, 0x1, 0x0, 0x0)
E:/Go/gopath/src/github.com/artonge/go-csv-tag/load.go:78 +0x10b