The control flow code blocks in Go are much like other popular programming languages, but there are also many differences. This article will show these similarities and differences.
if-else
two-way conditional execution block.
for
loop block.
switch-case
multi-way conditional execution block.
for-range
loop block for
container types.
type-switch
multi-way conditional execution block for
interface types.
select-case
block for
channel types.
Like many other popular languages, Go also supports break
,
continue
and goto
code execution jump statements.
Besides these, there is a special code jump statement in Go, fallthrough
.
Among the six kinds of control flow blocks,
except the if-else
control flow,
the other five are called breakable control flow blocks.
We can use break
statements to make executions
jump out of breakable control flow blocks.
for
and for-range
loop blocks are called
loop control flow blocks.
We can use continue
statements to end a loop step in advance
in a loop control flow block, i.e. continue to the next iteration of the loop.
Please note, each of the above mentioned control flow blocks is a statement, and it may contain many other sub-statements.
Above mentioned control flow statements are all the ones in narrow sense. The mechanisms introduced in the next article, goroutines, deferred function calls and panic/recover, and the concurrency synchronization techniques introduced in the later article concurrency synchronization overview can be viewed as control flow statements in broad sense.
Only the basic control flow code blocks and code jump statements will be explained in the current article, other ones will be explained in many other Go 101 articles later.
if-else
Control Flow Blocksif-else
code block is like
if InitSimpleStatement; Condition {
// do something
} else {
// do something
}
if
and else
are keywords.
Like many other programming languages, the else
branch is optional.
The InitSimpleStatement
portion is also optional.
It must be a simple statement if it is present.
If it is absent, we can view it as a blank statement (one kind of simple statements).
In practice, InitSimpleStatement
is often
a short variable declaration or a pure assignment.
A Condition
must be an expression which results to a boolean value.
The Condition
portion can be enclosed in a pair of ()
or not,
but it can't be enclosed together with the InitSimpleStatement
portion.
If the InitSimpleStatement
in a if-else
block is present,
it will be executed before executing other statements in the if-else
block.
If the InitSimpleStatement
is absent,
then the semicolon following it is optional.
Each if-else
control flow forms
one implicit code block, one if
branch explicit code block
and one optional else
branch code block.
The two branch code blocks are both nested in the implicit code block.
Upon execution, if Condition
expression results in true
,
then the if
branch block will get executed, otherwise,
the else
branch block will get executed.
package main
import (
"fmt"
"math/rand"
"time"
)
func main() {
rand.Seed(time.Now().UnixNano()) // needed before Go 1.20
if n := rand.Int(); n%2 == 0 {
fmt.Println(n, "is an even number.")
} else {
fmt.Println(n, "is an odd number.")
}
n := rand.Int() % 2 // this n is not the above n.
if n % 2 == 0 {
fmt.Println("An even number.")
}
if ; n % 2 != 0 {
fmt.Println("An odd number.")
}
}
If the InitSimpleStatement
in a if-else
code block is a short variable declaration,
then the declared variables will be viewed as being declared in the top nesting implicit code block of the if-else
code block.
An else
branch code block can be implicit if the corresponding else
is followed by another if-else
code block, otherwise, it must be explicit.
package main
import (
"fmt"
"time"
)
func main() {
if h := time.Now().Hour(); h < 12 {
fmt.Println("Now is AM time.")
} else if h > 19 {
fmt.Println("Now is evening time.")
} else {
fmt.Println("Now is afternoon time.")
h := h // the right one is declared above
// The just new declared "h" variable
// shadows the above same-name one.
_ = h
}
// h is not visible here.
}
for
Loop Control Flow Blocksfor
loop block is:
for InitSimpleStatement; Condition; PostSimpleStatement {
// do something
}
for
is a keyword.
The InitSimpleStatement
and PostSimpleStatement
portions
must be both simple statements, and the PostSimpleStatement
portion
must not be a short variable declaration.
Condition
must be an expression which result is a boolean value.
The three portions are all optional.
Unlike many other programming languages, the just mentioned three parts following
the for
keyword can't be enclosed in a pair of ()
.
Each for
control flow forms at least two code blocks,
one is implicit and one is explicit.
The explicit one is nested in the implicit one.
The InitSimpleStatement
in a for
loop block
will be executed (only once) before executing other statements in the for
loop block.
The Condition
expression will be evaluated at each loop step.
If the evaluation result is false
, then the loop will end.
Otherwise the body (a.k.a., the explicit code block) of the loop will get executed.
The PostSimpleStatement
will be executed at the end of each loop step.
for
loop example. The example will print the integers from
0
to 9
.
for i := 0; i < 10; i++ {
fmt.Println(i)
}
If the
InitSimpleStatement
and PostSimpleStatement
portions are both absent (just view them as blank statements),
their nearby two semicolons can be omitted.
The form is called as condition-only for
loop form.
It is the same as the while
loop in other languages.
var i = 0
for ; i < 10; {
fmt.Println(i)
i++
}
for i < 20 {
fmt.Println(i)
i++
}
If the
Condition
portion is absent,
compilers will view it as true
.
for i := 0; ; i++ { // <=> for i := 0; true; i++ {
if i >= 10 {
// "break" statement will be explained below.
break
}
fmt.Println(i)
}
// The following 4 endless loops are
// equivalent to each other.
for ; true; {
}
for true {
}
for ; ; {
}
for {
}
If the
InitSimpleStatement
in a for
block is a short variable declaration statement,
then the declared variables will be viewed as being declared in the top nesting implicit code block of the for
block.
For example, the following code snippet prints 012
instead of 0
.
for i := 0; i < 3; i++ {
fmt.Print(i)
// The left i is a new declared variable,
// and the right i is the loop variable.
i := i
// The new declared variable is modified, but
// the old one (the loop variable) is not yet.
i = 10
_ = i
}
A
break
statement can be used to make execution jump out of the
for
loop control flow block in advance,
if the for
loop control flow block is the innermost breakable
control flow block containing the break
statement.
i := 0
for {
if i >= 10 {
break
}
fmt.Println(i)
i++
}
A
continue
statement can be used to end the current loop step
in advance (PostSimpleStatement
will still get executed),
if the for
loop control flow block is the innermost
loop control flow block containing the continue
statement.
For example, the following code snippet will print 13579
.
for i := 0; i < 10; i++ {
if i % 2 == 0 {
continue
}
fmt.Print(i)
}
switch-case
Control Flow Blocks
switch-case
control flow block is
one kind of conditional execution control flow blocks.
switch-case
block is
switch InitSimpleStatement; CompareOperand0 {
case CompareOperandList1:
// do something
case CompareOperandList2:
// do something
...
case CompareOperandListN:
// do something
default:
// do something
}
In the full form,
switch
, case
and default
are three keywords.
InitSimpleStatement
portion must be a simple statement.
The CompareOperand0
portion is an expression which is viewed as
a typed value (if it is an untyped value,
then it is viewed as a type value of its default type),
hence it can't be an untyped nil
.
CompareOperand0
is called as switch expression in Go specification.
CompareOperandListX
(X
may represent from 1
to N
)
portions must be a comma separated expression list.
Each of these expressions shall be comparable with CompareOperand0
.
Each of these expressions is called as a case expression in Go specification.
If a case expression is an untyped value,
then it must be implicitly convertible to the type of the switch expression
in the same switch-case
control flow.
If the conversion is impossible to achieve, compilation fails.
Each case CompareOperandListX:
or default:
opens (and is followed by) an implicit code block.
The implicit code block and that case CompareOperandListX:
or default:
forms a branch.
Each such branch is optional to be present.
We call an implicit code block in such a branch as a branch code block later.
There can be at most one default
branch
in a switch-case
control flow block.
Besides the branch code blocks,
each switch-case
control flow forms two code blocks,
one is implicit and one is explicit.
The explicit one is nested in the implicit one.
All the branch code blocks are nested in the explicit one
(and nested in the implicit one indirectly).
switch-case
control flow blocks are breakable,
so break
statements can also be used in any branch code block
in a switch-case
control flow block to make execution jump out of
the switch-case
control flow block in advance.
The InitSimpleStatement
in a for
loop block
will be executed (only once) before executing other statements in the for
loop block.
The InitSimpleStatement
will get executed firstly
when a switch-case
control flow gets executed,
then the switch CompareOperand0
expression will be evaluated and only evaluated once.
The evaluation result is always a typed value.
The evaluation result will be compared (by using the ==
operator)
with the evaluation result of each case expression in the
CompareOperandListX
expression lists,
from top to down and from left to right.
If a case expression is found to be equal to CompareOperand0
,
the comparison process stops and the corresponding branch code block
of the expression will be executed.
If none case expressions are found to be equal to CompareOperand0
,
the default branch code block (if it is present) will get executed.
switch-case
control flow example:
package main
import (
"fmt"
"math/rand"
"time"
)
func main() {
rand.Seed(time.Now().UnixNano()) // needed before Go 1.20
switch n := rand.Intn(100); n%9 {
case 0:
fmt.Println(n, "is a multiple of 9.")
// Different from many other languages,
// in Go, the execution will automatically
// jumps out of the switch-case block at
// the end of each branch block.
// No "break" statement is needed here.
case 1, 2, 3:
fmt.Println(n, "mod 9 is 1, 2 or 3.")
// Here, this "break" statement is nonsense.
break
case 4, 5, 6:
fmt.Println(n, "mod 9 is 4, 5 or 6.")
// case 6, 7, 8:
// The above case line might fail to compile,
// for 6 is duplicate with the 6 in the last
// case. The behavior is compiler dependent.
default:
fmt.Println(n, "mod 9 is 7 or 8.")
}
}
The rand.Intn
function returns a non-negative int
random value which is smaller than the specified argument.
Note, if any two case expressions in a switch-case
control flow
can be detected to be equal at compile time, then a compiler may reject the latter one.
For example, the standard Go compiler thinks the case 6, 7, 8
line
in the above example is invalid if that line is not commented out.
But other compilers may think that line is okay.
In fact, the current standard Go compiler (version 1.20)
allows duplicate boolean case expressions,
and gccgo (v8.2) allows both duplicate boolean and string case expressions.
switch-case
control block.
Then how to let the execution slip into the next branch code block?
Go provides a fallthrough
keyword to do this task.
For example, in the following example, every branch code block
will get executed, by their orders, from top to down.
rand.Seed(time.Now().UnixNano()) // needed before Go 1.20
switch n := rand.Intn(100) % 5; n {
case 0, 1, 2, 3, 4:
fmt.Println("n =", n)
// The "fallthrough" statement makes the
// execution slip into the next branch.
fallthrough
case 5, 6, 7, 8:
// A new declared variable also called "n",
// it is only visible in the current
// branch code block.
n := 99
fmt.Println("n =", n) // 99
fallthrough
default:
// This "n" is the switch expression "n".
fmt.Println("n =", n)
}
Please note,
fallthrough
statement must be the final statement in a branch.
fallthrough
statement can't show up
in the final branch in a switch-case
control flow block.
fallthrough
uses are all illegal.
switch n := rand.Intn(100) % 5; n {
case 0, 1, 2, 3, 4:
fmt.Println("n =", n)
// The if-block, not the fallthrough statement,
// is the final statement in this branch.
if true {
fallthrough // error: not the final statement
}
case 5, 6, 7, 8:
n := 99
fallthrough // error: not the final statement
_ = n
default:
fmt.Println(n)
fallthrough // error: show up in the final branch
}
The InitSimpleStatement
and CompareOperand0
portions
in a switch-case
control flow are both optional.
If the CompareOperand0
portion is absent,
it will be viewed as true
,
a typed value of the built-in type bool
.
If the InitSimpleStatement
portion is absent,
the semicolon following it can be omitted.
switch n := 5; n {
}
switch 5 {
}
switch _ = 5; {
}
switch {
}
For the latter two
switch-case
control flow blocks in the last example,
as above has mentioned, each of the absent CompareOperand0
portions is viewed as a typed value true
of the built-in type bool
.
So the following code snippet will print hello
.
switch { // <=> switch true {
case true: fmt.Println("hello")
default: fmt.Println("bye")
}
Another obvious difference from many other languages is the order of the
default
branch
in a switch-case
control flow block can be arbitrary.
For example, the following three switch-case
control flow blocks
are equivalent to each other.
switch n := rand.Intn(3); n {
case 0: fmt.Println("n == 0")
case 1: fmt.Println("n == 1")
default: fmt.Println("n == 2")
}
switch n := rand.Intn(3); n {
default: fmt.Println("n == 2")
case 0: fmt.Println("n == 0")
case 1: fmt.Println("n == 1")
}
switch n := rand.Intn(3); n {
case 0: fmt.Println("n == 0")
default: fmt.Println("n == 2")
case 1: fmt.Println("n == 1")
}
goto
Statement and Label Declaration
Like many other languages, Go also supports goto
statement.
A goto
keyword must be followed by a label to form a statement.
A label is declared with the form LabelName:
,
where LabelName
must be an identifier.
A label which name is not the blank identifier must be used at least once.
A goto
statement will make the execution jump to the next statement
following the declaration of the label used in the goto
statement.
So a label declaration must be followed by one statement.
A label must be declared within a function body. A use of a label can appear before or after the declaration of the label. But a label is not visible (and can't appear) outside the innermost code block the label is declared in.
The following example uses agoto
statement and a label to implement a loop control flow.
package main
import "fmt"
func main() {
i := 0
Next: // here, a label is declared.
fmt.Println(i)
i++
if i < 5 {
goto Next // execution jumps
}
}
As mentioned above, a label is not visible (and can't appear) outside the innermost code block the label is declared in. So the following example fails to compile.
package main
func main() {
goto Label1 // error
{
Label1:
goto Label2 // error
}
{
Label2:
}
}
Note that, if a label is declared within the scope of a variable, then the uses of the label can't appear before the declaration of the variable. Identifier scopes will be explained in the article blocks and scopes in Go later.
The following example also fails to compile.package main
import "fmt"
func main() {
i := 0
Next:
if i >= 5 {
// error: jumps over declaration of k
goto Exit
}
k := i + i
fmt.Println(k)
i++
goto Next
// This label is declared in the scope of k,
// but its use is outside of the scope of k.
Exit:
}
The just mentioned rule may change later.
Currently, to make the above code compile okay,
we must adjust the scope of the variable k
.
There are two ways to fix the problem in the last example.
k
.
func main() {
i := 0
Next:
if i >= 5 {
goto Exit
}
// Create an explicit code block to
// shrink the scope of k.
{
k := i + i
fmt.Println(k)
}
i++
goto Next
Exit:
}
The other way is to enlarge the scope of the variable k
.
func main() {
var k int // move the declaration of k here.
i := 0
Next:
if i >= 5 {
goto Exit
}
k = i + i
fmt.Println(k)
i++
goto Next
Exit:
}
break
and continue
Statements With Labels
A goto
statement must contain a label.
A break
or continue
statement can also
contain a label, but the label is optional.
Generally, break
containing labels are used in nested breakable control flow blocks
and continue
statements containing labels are used in nested loop control flow blocks.
If a break
statement contains a label,
the label must be declared just before a breakable control flow block
which contains the break
statement.
We can view the label name as the name of the breakable control flow block.
The break
statement will make execution jump out of
the breakable control flow block, even if the breakable control flow block
is not the innermost breakable control flow block containing
break
statement.
If a continue
statement contains a label,
the label must be declared just before a loop control flow block
which contains the continue
statement.
We can view the label name as the name of the loop control flow block.
The continue
statement will end the current loop step of
the loop control flow block in advance, even if the loop control flow block
is not the innermost loop control flow block containing
the continue
statement.
break
and continue
statements with labels.
package main
import "fmt"
func FindSmallestPrimeLargerThan(n int) int {
Outer:
for n++; ; n++{
for i := 2; ; i++ {
switch {
case i * i > n:
break Outer
case n % i == 0:
continue Outer
}
}
}
return n
}
func main() {
for i := 90; i < 100; i++ {
n := FindSmallestPrimeLargerThan(i)
fmt.Print("The smallest prime number larger than ")
fmt.Println(i, "is", n)
}
}
The Go 101 프로젝트는 Github 에서 호스팅됩니다. 오타, 문법 오류, 부정확한 표현, 설명 결함, 코드 버그, 끊어진 링크와 같은 모든 종류의 실수에 대한 수정 사항을 제출하여 Go 101을 개선을 돕는 것은 언제나 환영합니다.
주기적으로 Go에 대한 깊이 있는 정보를 얻고 싶다면 Go 101의 공식 트위터 계정인 @go100and1을 팔로우하거나 Go 101 슬랙 채널에j가입해주세요.
reflect
표준 패키지sync
표준 패키지sync/atomic
표준 패키지