CODEWITHSUNDEEP
arraysperformanceloopsgoslice
juancsr
juancsr

Reputation: 329

an efficient way to loop an slice/array in go

basically I have this:

package main

import "fmt"

type Struct1 struct {
    id   int
    name string
}

type Struct2 struct {
    id       int
    lastname string
}

type Struct3 struct {
    id   int
    real bool
}

func main() {
    var (
     s1 []Struct1
     s2 []Struct2
     s3 []Struct3
    )
    s1 = append(s1, Struct1{id: 1, name: "Eliot"}, Struct1{id: 2, name: "Tyrell"}, Struct1{id: 3, name: "Mr Robot"})
    s2 = append(s2, Struct2{id: 1, lastname: "Anderson"}, Struct2{id: 2, lastname: "Wellick"})
    s3 = append(s3, Struct3{id: 1, real: true}, Struct3{id: 2, real: true}, Struct3{id: 3, real: false})
}

I want to show something like this:

But I don't want to loop the s1 inside the s2 and then inside the s3

Example:

for i := 0; i < len(s1); i++ {
        for j := 0; j < len(s2); j++ {
            if s1[i].id == s2[j].id {
                for k := 0; k < len(s3); k++ {
                    if s2[j].id == s3[k].id {
                        // some code ...
                    }
                }
            }
        }
    }

So, what other ways are there to doing that?

Upvotes: 1

Views: 3367

Answers (3)

Eugene Lisitsky
Eugene Lisitsky

Reputation: 12845

In fact to join all first elements, then all second etc, you don’t need to make a cycle in a cycle:

Prepare every slice as a map:

m1 := make(map[int]string)
for i:=0; i < len(s1); i ++ {
    m1[s1[i].id] = s1[i].Name
}

The same for 2 other slices. Finally iterate once:

for i:=0; i < len(s1); i ++ {
    fmt.Println(m1[i], m2[i], m3[i])
}

This solution supposes all slices have corresponding items. If not, you should decide what to do with an element without its fragment: ignore, replace blank with some placeholders, break whole cycle etc. For example, you decided firstName is mandatory (we will iterate by it), secondName is optional and should be replaced with ? in case of absence, and real is absolutely obligatory for the whole cycle - if it's absent we break further work and return empty slice:

type Person struct { // corrected struct from the neighbour recipe
     firstName: string
     lastName: string 
     real: bool
}

var persons []Person // slice for result

for i, firstName := range m1 { // iterate by keys and values of m1
    secondName, ok := m2[i]
    if !ok {
       secondName = "?"
    }
    real, ok := m3[i]
    if !ok {
       persons = []Person
       break
    }
    person := Person{firstName, secondName, real}
    persons = append(persons, person)
}

Upvotes: 0

Ravi R
Ravi R

Reputation: 1782

The right way would be to put them in a hash(called map in Golang). That way you can get performance and you could do with only one loop iterating over id's.

Here's an example with your sample data:

package main

import (
    "fmt"
)

type Struct1 struct {
    id   int
    name string
}

type Struct2 struct {
    id       int
    lastname string
}

type Struct3 struct {
    id   int
    real bool
}

func main() {
    //var (
    //s1 []Struct1
    //      s2 []Struct2
    //  s3 []Struct3
    //  )
    s1Hash := make(map[int]Struct1)
    s2Hash := make(map[int]Struct2)
    s3Hash := make(map[int]Struct3)

    s11 := Struct1{id: 1, name: "Eliot"}
    s12 := Struct1{id: 2, name: "Tyrell"}
    s13 := Struct1{id: 3, name: "Mr Robot"}
    s1Hash[s11.id] = s11
    s1Hash[s12.id] = s12
    s1Hash[s13.id] = s13

    s21 := Struct2{id: 1, lastname: "Anderson"}
    s22 := Struct2{id: 2, lastname: "Wellick"}
    s2Hash[s21.id] = s21
    s2Hash[s22.id] = s22

    s31 := Struct3{id: 1, real: true}
    s32 := Struct3{id: 2, real: true}
    s33 := Struct3{id: 3, real: false}
    s3Hash[s31.id] = s31
    s3Hash[s32.id] = s32
    s3Hash[s33.id] = s33

    //s1 = append(s1, Struct1{id: 1, name: "Eliot"}, Struct1{id: 2, name: "Tyrell"}, Struct1{id: 3, name: "Mr Robot"})
    //s2 = append(s2, Struct2{id: 1, lastname: "Anderson"}, Struct2{id: 2, lastname: "Wellick"})
    //s3 = append(s3, Struct3{id: 1, real: true}, Struct3{id: 2, real: true}, Struct3{id: 3, real: false})

    //i to loop over possible id range
    for i := 1; i <= len(s1Hash); i++ {
        fmt.Println("i is ", i)
        if _, ok := s1Hash[i]; ok {
            fmt.Printf("Name: %s ", s1Hash[i].name)
        }

        if _, ok := s2Hash[i]; ok {
            fmt.Printf(" Lastname: %s ", s2Hash[i].lastname)
        }

        if _, ok := s3Hash[i]; ok {
            fmt.Printf(" Real: %t\n", s3Hash[i].real)
        }
        //fmt.Printf("%s %s real:%t\n", s1Hash[i].name, s2[i].lastname, s3[i].real)
    }

}

Output:

i is  1
Name: Eliot  Lastname: Anderson  Real: true
i is  2
Name: Tyrell  Lastname: Wellick  Real: true
i is  3
Name: Mr Robot  Real: false

Check this out on playground. Hope this helps!

p.s. : Eventually if you may delete all struct entries for some ID's and add newer ID's - you can consider adding the valid ID's into a map map[int]bool (mymap[id] = true) and iterate over the map using range instead of the for i.. as above.

Upvotes: 2

Daniel Tran
Daniel Tran

Reputation: 6171

Use a map for indexing the data by the person name.

type Person struct {
     firstName: string
     lastName: string 
     real: string
}

var data map[int]Person

Then look up and put data to the map.

Upvotes: 0

Related Questions