Just getting started writing Go code and I ran into an interesting problem.
Is there a way to easily iterate over the items in an array that is brought in as an empty interface without code duplication? Consider the following:
function(someArr interface{}){
switch someArr.(type){
case []int :
arr := (someArr).([]int)
for i := range (arr) {
// CODE
}
case []string :
arr := (someArr).([]string)
for i := range (arr) {
// CODE
}
}
}
In this example the code in CODE is exactly the same. However, I cannot take it out of the switch because the type assertion arr would fall out of scope. Similarly, I can't define arr before the switch because I don't know what type it will be. It's possible that this just cannot be done. In that case, what's a better idiom for this kind of thing when I'm, say, parsing JSON with an irregular schema (some arrays of ints, some arrays or strings)?
Your example is not idiomatic Go code, even though the idiomatic one lexically seems violating the DRY principle as well.
The key point to understand is that 'x' is a separate, differently typed variable in each type case:
function(someArr interface{}){
switch x := someArr.(type) {
case []int:
for i := range x {
// CODE
}
case []string:
for i := range x {
// CODE
}
}
}
You can use the reflect package to iterate over arbitrary slices. But implementing the special cases (like []int) explicitly is generally faster and is often done in addition to avoid reflection in common cases.
package main
import "fmt"
import "reflect"
func foo(values interface{}) {
rv := reflect.ValueOf(values)
if rv.Kind() != reflect.Slice {
return
}
n := rv.Len()
for i := 0; i < n; i++ {
value := rv.Index(i).Interface()
fmt.Println(value)
}
}
func main() {
foo([]int{1, 3, 3, 7})
}
Edit: I'm not sure why somebody has down voted the question and my answer, but there are cases where you need to deal with code like that. Even the standard library contains plenty of it, take a look at "fmt", "gob", "json", "xml" and "template" for example. The questioner might face a similar problem.
Related
I decided that now that generics have been introduced into Go that something like map/reduce should be possible. So, I took a naive stab at it and I get the error:
./prog.go:18:36: cannot use thing (variable of type int) as type I in argument to mapper
Which doesn't explain if the problem is fundamental or I am simply doing something wrong syntactically. Can generic map/reduce be implemented in Go?
package main
import "fmt"
func main() {
things := []int{1, 2, 3, 4}
results := Map(things, func(t int) int {
return t + 1
})
fmt.Printf("%v", results)
}
func Map[I interface{}, O interface{}](things []I, mapper func(thing I) O) []O {
results := make([]O, 0, len(things))
for thing := range things {
results = append(results, mapper(thing))
}
return results
}
You have incorrect use of range. A single variable extracted from range will be the index (type int), not the value (type I, which is only coincidentally int in this case).
Try
for _, thing := range things{...}
This can be done quite easily. You have an error in your code, though right here:
for thing := range things {
You are iterating over the index values (int), not the values of type I. You're also specifying 2 constraints (types I and O) both set to be interface{}. You can just use any instead (it's shorthand for interface{})
So simply write:
func Map[T any, O any](things []T, mapper func(thing T) O) []O {
result := make([]O, 0, len(things))
for _, thing := range things {
result = append(result, mapper(thing))
}
return result
}
Demo
This is quite closely related to some code I reviewed on codereview exchange here. After going through the code, and writing snippets with a ton of suggestions, I decided to just create a package and throw it up on github instead. You can find the repo here.
In it, there's some examples that may come in handy, or help you work through some other quirks WRT generics in golang. I wsa specifically thinking about this bit, where you can filter a generic map type using callbacks like so:
// given the sMap type
type sMap[K comparable, V any] struct {
mu *sync.RWMutex
m map[K]V
}
// Filter returns a map containing the elements that matched the filter callback argument
func (s *sMap[K, V]) Filter(cb func(K, V) bool) map[K]V {
s.mu.RLock()
defer s.mu.RUnlock()
ret := make(map[K]V, len(s.m))
for k, v := range s.m {
if cb(k, v) {
ret[k] = v
}
}
return ret
}
I have two maps, both of them are keyed by strings, but the values are of two different custom types.
map[string]type1
map[string]type2
Now I want to write a function which can take an argument of either of these two types, because that function only looks at the keys and doesn't care about the values at all. So I think it should look like this:
func takeTheMap(argument map[string]interface{}) {
...
But that doesn't work due to:
cannot use myVariable (type map[string]customType) as type map[string]interface {} in argument to takeTheMap
https://play.golang.org/p/4Xkhi4HekO5
Can I make that work somehow?
The only polymorphism in Go is interfaces. The only alternatives to that are reflection, duplication, or rethinking the broader design so that you don't need to do what you're trying to do here.
If the last option isn't a possibility, personally I would recommend duplication, since it's a whole four lines of code.
keys := make([]string, 0, len(myMap))
for key,_ := range myMap {
keys = append(keys,key)
}
A big complicated generic helper seems kind of unnecessary.
A solution using an interface. This example may seem a bit overkill and it may be better to in your case (I'm not sure, not enough details in your example) to just use a couple of for loops.
package main
import (
"fmt"
)
type foo bool
type bar string
type mapOne map[string]foo
type mapTwo map[string]bar
func (m mapOne) Keys() []string {
s := []string{}
for k := range m {
s = append(s, k)
}
return s
}
func (m mapTwo) Keys() []string {
s := []string{}
for k := range m {
s = append(s, k)
}
return s
}
type ToKeys interface {
Keys() []string
}
func main() {
m1 := mapOne{"one": true, "two": false}
m2 := mapTwo{"three": "foo", "four": "bar"}
doSomething(m1)
doSomething(m2)
}
func doSomething(m ToKeys) {
fmt.Println(m.Keys())
}
Playground example
Cannot Range Over List Type Interface {} In Function Using Go.
for me is important then i execute for in a function.
How can fix?
package main
import (
"fmt"
)
type MyBoxItem struct {
Name string
}
type MyBox struct {
Items []MyBoxItem
}
func (box *MyBox) AddItem(item MyBoxItem) []MyBoxItem {
box.Items = append(box.Items, item)
return box.Items
}
func PrintCustomArray(list interface{}) interface{} {
//items := reflect.ValueOf(list)
for _, v := range list {
fmt.Println(v.Key,v.Value)
}
return 0
}
func main() {
items := []MyBoxItem{}
item := MyBoxItem{Name: "Test Item 1"}
box := MyBox{items}
box.AddItem(item)
fmt.Println((box.Items))
PrintCustomArray(box.Items)
}
https://play.golang.org/p/ZcIBLMliq3
Error : cannot range over list (type interface {})
How can fix?
Note
The answer below describes, in broad strokes, 2 possible approaches: using interfaces, and using specific types. The approach focusing on interfaces is mentioned for completeness sake. IMHO, the case you've presented is not a viable use-case for interfaces.
Below, you'll find a link to a playground example that uses both techniques. It should be apparent to anyone that the interface approach is too cumbersome if for this specific case.
Quite apart from the fact that you don't really seem to be too familiar with how loops work in go (v.Key and v.Value are non-existent fields for example), I'll attempt to answer your question.
You are passing a list to your function, sure enough, but it's being handled as an interface{} type. That means your function accepts, essentially, any value as an argument. You can't simply iterate over them.
What you can do is use type assertions to convert the argument to a slice, then another assertion to use it as another, specific interface:
type Item interface{
key() string
val() string
}
func (i MyBoxItem) key() string {
return i.Key
}
func (i MyBoxItem) val() string {
return i.Value
}
func PrintCustomArray(list interface{}) error {
listSlice, ok := list.([]interface{})
if !ok {
return fmt.Errorf("Argument is not a slice")
}
for _, v := range listSlice {
item, ok := v.(Item)
if !ok {
return fmt.Errorf("element in slice does not implement the Item interface")
}
fmt.Println(item.key(), item.val())
}
return nil
}
But let's be honest, a function like this only works if a slice is passed as an argument. So having that first type assertion in there makes no sense whatsoever. At the very least, changing the function to something like this makes a lot more sense:
func PrintCustomArray(list []interface{})
Then, because we're not expecting an array as such, but rather a slice, the name should be changed to PrintCustomSlice.
Lastly, because we're using the same type assertion for every value in the slice, we might as well change the function even more:
// at this point, we'll always return 0, which is pointless
// just don't return anything
func PrintCustomSlice(list []Item) {
for _, v := range list {
fmt.Println(v.key(), v.val())
}
}
The advantages of a function like this is that it can still handle multiple types (all you have to do is implement the interface). You don't need any kind of expensive operations (like reflection), or type assertions.
Type assertions are very useful, but in a case like this, they merely serve to hide problems that would otherwise have resulted in a compile-time error. Go's interface{} type is a very useful thing, but you seem to be using it to get around the type system. If that's what you want to achieve, why use a typed language in the first place?
Some closing thoughts/remarks: If your function is only going to be used to iterate over specific "thing", you don't need the interfaces at all, simply specify the type you're expecting to be passed to the function in the first place. In this case that would be:
func PrintCustomSlice(list []MyBoxItem) {
for _, v := range list {
fmt.Println(v.Key, v.Value)
}
}
Another thing that I've noticed is that you seem to be exporting everything (all functions, types, and fields start with a capital letter). This, in go, is considered bad form. Only export what needs to be public. In the main package, that usually means you're hardly export anything.
Lastly, as I mentioned at the start: you don't seem to have a firm grasp on the basics just yet. I'd strongly recommend you go through the interactive tour. It covers the basics nicely, but shows you the features of the language at a decent pace. It doesn't take long, and is well worth taking a couple of hours to complete
Playground demo
It's possible to implement PrintCustomArray using the reflect package, but most experienced Go programmers will write a simple for loop:
for _, i := range box.Items {
fmt.Println("Name:", i.Name)
}
https://play.golang.org/p/RhubiCpry0
You can also encapsulate it in a function:
func PrintCustomArray(items []MyBoxItem) {
for _, i := range items {
fmt.Println("Name:", i.Name)
}
}
https://play.golang.org/p/c4EPQIx1AH
Here since you are returning box.Items from AddItem(), Items is of the type []MyBoxItem , so list should be of type []MyBoxItem .Moreover you are returning 0 in PrintCustomArray and the return type you have set is {}interface.
func PrintCustomArray(list []MyBoxItem) {
//items := reflect.ValueOf(list)
for i, v := range list {
fmt.Println(i, v)
}
//return 0
}
Again, MyBoxItem struct has only one variable named Name so v.key v.value won't make any sense.
This is what the proper code should look like https://play.golang.org/p/ILoUwEWv6Y .
You need to clear your understanding about interfaces in go. This might help https://golang.org/doc/effective_go.html#interfaces_and_types .
I've got a handful of interfaces, and n number of structs that arbitrarily implement these interfaces. I'd like to keep an array of types and be able to run a loop over them to see which ones are implemented. Is it possible to store a type like this? I spent a little bit of time with the reflect package, but couldn't really find what I was looking for, I understand if maybe this isn't best practice. Trying to do something similar to this.. without a giant type switch, fallthrough, or if.. if... if.
type InterOne interface {
InterOneMethod() string
}
var interfaceMap = map[string]type {
"One": InterOne,
...
}
func doesHandle(any interface{}) []string {
var handles []string
for k, v := range interfaceMap {
if _, ok := any.(v); ok {
handles = append(handles, k)
}
}
return handles
}
EDIT: The answer marked as correct is technically right. I found that due to the comment about the method calling & the overuse of reflection, that this approach was a bad idea. Instead I went with a type switch to check for a single interface because fallthrough is not supported on type switch, and a large if.. if.. if.. with type assertions to be able to make the appropriate calls.
You can use reflect, notice that to get the type of an interface the only way is to use reflect.TypeOf((*INTERFACE)(nil)).Elem(), here's a working example:
var interfaceMap = map[string]reflect.Type{
"One": reflect.TypeOf((*InterOne)(nil)).Elem(),
....
}
func doesHandle(any interface{}) []string {
t := reflect.TypeOf(any)
var handles []string
for k, v := range interfaceMap {
if t.Implements(v) {
handles = append(handles, k)
}
}
return handles
}
playground
Is there a way to write a generic array/slice deduplication in go, for []int we can have something like (from http://rosettacode.org/wiki/Remove_duplicate_elements#Go ):
func uniq(list []int) []int {
unique_set := make(map[int] bool, len(list))
for _, x := range list {
unique_set[x] = true
}
result := make([]int, len(unique_set))
i := 0
for x := range unique_set {
result[i] = x
i++
}
return result
}
But is there a way to extend it to support any array? with a signature like:
func deduplicate(a []interface{}) []interface{}
I know that you can write that function with that signature, but then you can't actually use it on []int, you need to create a []interface{} put everything from the []int into it, pass it to the function then get it back and put it into a []interface{} and go through this new array and put everything in a new []int.
My question is, is there a better way to do this?
While VonC's answer probably does the closest to what you really want, the only real way to do it in native Go without gen is to define an interface
type IDList interface {
// Returns the id of the element at i
ID(i int) int
// Returns the element
// with the given id
GetByID(id int) interface{}
Len() int
// Adds the element to the list
Insert(interface{})
}
// Puts the deduplicated list in dst
func Deduplicate(dst, list IDList) {
intList := make([]int, list.Len())
for i := range intList {
intList[i] = list.ID(i)
}
uniques := uniq(intList)
for _,el := range uniques {
dst.Insert(list.GetByID(el))
}
}
Where uniq is the function from your OP.
This is just one possible example, and there are probably much better ones, but in general mapping each element to a unique "==able" ID and either constructing a new list or culling based on the deduplication of the IDs is probably the most intuitive way.
An alternate solution is to take in an []IDer where the IDer interface is just ID() int. However, that means that user code has to create the []IDer list and copy all the elements into that list, which is a bit ugly. It's cleaner for the user to wrap the list as an ID list rather than copy, but it's a similar amount of work either way.
The only way I have seen that implemented in Go is with the clipperhouse/gen project,
gen is an attempt to bring some generics-like functionality to Go, with some inspiration from C#’s Linq and JavaScript’s underscore libraries
See this test:
// Distinct returns a new Thing1s slice whose elements are unique. See: http://clipperhouse.github.io/gen/#Distinct
func (rcv Thing1s) Distinct() (result Thing1s) {
appended := make(map[Thing1]bool)
for _, v := range rcv {
if !appended[v] {
result = append(result, v)
appended[v] = true
}
}
return result
}
But, as explained in clipperhouse.github.io/gen/:
gen generates code for your types, at development time, using the command line.
gen is not an import; the generated source becomes part of your project and takes no external dependencies.
You could do something close to this via an interface. Define an interface, say "DeDupable" requiring a func, say, UniqId() []byte, which you could then use to do the removing of dups. and your uniq func would take a []DeDupable and work on it