yogofn

Functional Map/Reduce/Filter for Golang

install

go get github.com/johncgriffin/yogofn

Yogofn offers a few slice generating operations as alternatives to Go's for-loops.

For instance, let's find the max daily temperature range given two parallel lists of low and high daily temperatures. Here's standard Go code given two float64 slices.

var maxDailyRange float64
for i:=0; i<len(highs) && i<len(lows); i++ {
    maxDailyRange = math.Max(highs[i] - lows[i], maxDailyRange)
}

And here's using yogofn:

maxDailyRange := Reduce(MaxF64, Map(SubF64, highs, lows)).(float64)

Notice that the final outside operation (Reduce in this instance) required a cast to inform Go of the final type. Inputs are either arrays or slices of any type. Yogofn offers:

• Map(f,slice(s)) -> slice projection
• Filter(f,slice) -> slice selection
• Reduce(f,slice(s)) -> scalar
• Every(f,slice) -> bool
• Any(f,slice) -> bool

Notice that because Reduce and Map can take more than one list, Zip is effected via

zipped := Map(func(x,y string) []string { return []string{x,y} }, xs, ys).([][]string)

A companion package is yogofn/reducers which contains simple but common scalar reductions for int and float64 numbers.

Reflection Performance

Not suprisingly, reflection slows down standard Go about 200 times. Using Go's type switch, a few optimizations were placed into Map/Filter/Reduce for common data types float64, int, and string. Those operations are implemented as normal typed Go loops.