slices2

Mapping over a slice, including the case where the mapping can fail.

The standard library has slices for searching, sorting and comparing, but no Map. Writing the loop by hand is three lines, and two of them are the ones people get wrong: the result slice is allocated without capacity, or the error from element seventeen is dropped because the loop had nowhere to put it.

import "gitlab.internal/tuan.quang.tran/aioz-common/slices2"

Usage

ids := slices2.Map(workers, func(w Worker) string { return w.ID })

// Convert stops at the first failure and returns it.
addrs, err := slices2.Convert(raw, netip.ParseAddr)
if err != nil {
    return err
}

// ConvertErrs keeps going and returns every failure, so a bad batch reports
// all of its bad rows rather than only the first.
rows, errs := slices2.ConvertErrs(lines, parseRow)

Choosing between Convert and ConvertErrs

Convert is for a pipeline: the first bad element makes the whole result meaningless, so there is nothing to gain from parsing the rest.

ConvertErrs is for a batch someone has to fix. Reporting one broken row at a time turns a single correction pass into as many passes as there are broken rows, which is the thing that makes an import job miserable.

API

func Map[In, Out any](xs []In, fn func(In) Out) []Out
func Convert[In, Out any](xs []In, fn func(In) (Out, error)) ([]Out, error)
func ConvertErrs[In, Out any](xs []In, fn func(In) (Out, error)) ([]Out, []error)