Error Handling
Welcome to tutorial no. 30 in Golang tutorial series.
What are errors?
Errors indicate any abnormal condition occurring in the program. Let’s say we are trying to open a file and the file does not exist in the file system. This is an abnormal condition and it’s represented as an error.
Errors in Go are plain old values. Just like any other built-in type such as int, float64, … error values can be stored in variables, passed as parameters to functions, returned from functions, and so on.
Errors are represented using the built-in error type. We will learn more about the error type later in this tutorial.
Example
Let’s start right away with an example program that tries to open a file that does not exist.
1package main
2
3import (
4 "fmt"
5 "os"
6)
7
8func main() {
9 f, err := os.Open("/test.txt")
10 if err != nil {
11 fmt.Println(err)
12 return
13 }
14 fmt.Println(f.Name(), "opened successfully")
15}
In line no. 9 of the program above, we are trying to open the file at path /test.txt(which will obviously not exist in the playground). The Open function of the os package has the following signature,
*func Open(name string) (File, error)
If the file has been opened successfully, then the Open function will return the file handler and error will be nil. If there is an error while opening the file, a non-nil error will be returned.
If a function or method returns an error, then by convention it has to be the last value returned from the function. Hence the Open function returns error as the last value.
The idiomatic way of handling errors in Go is to compare the returned error to nil. A nil value indicates that no error has occurred and a non-nil value indicates the presence of an error. In our case, we check whether the error is not nil in line no. 10. If it is not nil, we simply print the error and return from the main function.
Running this program will print
open /test.txt: No such file or directory
Perfect 😃. We get an error stating that the file does not exist.
Error type representation
Let’s dig a little deeper and see how the built in error type is defined. error is an interface type with the following definition,
1type error interface {
2 Error() string
3}
It contains a single method with the signature Error() string. Any type which implements this interface can be used as an error. This method provides the description of the error.
When printing the error, fmt.Println function calls the Error() string method internally to get the description of the error. This is how the error description was printed in line no. 11 of the above sample program.
Different ways to extract more information from the error
Now that we know error is an interface type, let’s see how we can extract more information about an error.
In the example we saw above, we have just printed the description of the error. What if we wanted the actual path of the file which caused the error. One possible way to get this is to parse the error string. This was the output of our program,
open /test.txt: No such file or directory
We can parse this error message and get the file path “/test.txt” of the file which caused the error, but this is a dirty way of doing it. The error description can change at any time in newer versions of Go and our code will break.
Is there a better way to get the file name 🤔? The answer is yes, it can be done and the Go standard library uses different ways to provide more information about errors. Let’s look at them one by one.
1. Converting the error to the underlying type and retrieving more information from the struct fields
If you read the documentation of the Open function carefully, you can see that it returns an error of type *PathError. PathError is a struct type and its implementation in the standard library is as follows,
1type PathError struct {
2 Op string
3 Path string
4 Err error
5}
6
7func (e *PathError) Error() string { return e.Op + " " + e.Path + ": " + e.Err.Error() }
In case you are interested to know where the above source code exists, it can be found here https://cs.opensource.google/go/go/+/refs/tags/go1.19:src/io/fs/fs.go;l=250
From the above code, you can understand that *PathError implements the error interface by declaring the Error() string method. This method concatenates the operation, path, and the actual error and returns it. Thus we got the error message,
open /test.txt: No such file or directory
The Path field of PathError struct contains the path of the file which caused the error.
We can use the As function from errors package to convert the error to its underlying type. The As function’s description talks about error chain. Please ignore it for now. We will understand how error chain and wrapping works in a separate tutorial.
A simple description of As is that it tries to convert the error to a error type and returns either true or false indicating whether the conversion is successful or not.
A program will make things clear. Let’s modify the program we wrote above and print the path using the As function.
1package main
2
3import (
4 "errors"
5 "fmt"
6 "os"
7)
8
9func main() {
10 f, err := os.Open("test.txt")
11 if err != nil {
12 var pErr *os.PathError
13 if errors.As(err, &pErr) {
14 fmt.Println("Failed to open file at path", pErr.Path)
15 return
16 }
17 fmt.Println("Generic error", err)
18 return
19 }
20 fmt.Println(f.Name(), "opened successfully")
21}
In the above program, we first check whether the error is not nil in line no. 11 and then we use the As function in line no. 13 to convert err to *os.PathError. If the conversion is successful, As will return true. Then we print the path using pErr.Path in line no. 14.
If you are wondering why pErr is a pointer, the reason is, the error interface is implemented by the pointer of PathError and hence pErr is a pointer. The below code shows that *PathError implements the error interface.
func (e *PathError) Error() string { return e.Op + " " + e.Path + ": " + e.Err.Error() }
The As function requires the second argument to be a pointer to the type that implements the error. Hence we pass &perr.
This program outputs,
Failed to open file at path test.txt
In case the underlying error is not of type *os.PathError, the control will reach line no. 17 and a generic error message will be printed.
Great 😃. We have successfully used the As function to get the file path from the error.
2. Retrieving more information using methods
The second way to get more information from the error is to find out the underlying type and get more information by calling methods on the struct type.
Let’s understand this better by means of an example.
The DNSError struct type in the standard library is defined as follows,
1
2type DNSError struct {
3 ...
4}
5
6func (e *DNSError) Error() string {
7 ...
8}
9func (e *DNSError) Timeout() bool {
10 ...
11}
12func (e *DNSError) Temporary() bool {
13 ...
14}
The DNSError struct has two methods Timeout() bool and Temporary() bool which return a boolean value that indicates whether the error is because of a timeout or is it a temporary one.
Let’s write a program that converts the error to *DNSError type and calls the above mentioned methods to determine whether the error is temporary or due to timeout.
1package main
2
3import (
4 "errors"
5 "fmt"
6 "net"
7)
8
9func main() {
10 addr, err := net.LookupHost("golangbot123.com")
11 if err != nil {
12 var dnsErr *net.DNSError
13 if errors.As(err, &dnsErr) {
14 if dnsErr.Timeout() {
15 fmt.Println("operation timed out")
16 return
17 }
18 if dnsErr.Temporary() {
19 fmt.Println("temporary error")
20 return
21 }
22 fmt.Println("Generic DNS error", err)
23 return
24 }
25 fmt.Println("Generic error", err)
26 return
27 }
28 fmt.Println(addr)
29}
Note: DNS lookups do not work in the playground. Please run this program in your local machine.
In the program above, in line no. 9, we are trying to get the IP address of an invalid domain name golangbot123.com. In line no. 13 we get the underlying value of the error by using the As function and converting it to *net.DNSError. Then we check whether the error is due to timeout or is temporary in line nos. 14 and 18 respectively.
In our case, the error is neither temporary nor due to timeout and hence the program will print,
Generic DNS error lookup golangbot123.com: no such host
If the error was temporary or due to a timeout, then the corresponding if statement would have executed and we can handle it appropriately.
3. Direct comparison
The third way to get more details about an error is the direct comparison with a variable of type error. Let’s understand this by means of an example.
The Glob function of the filepath package is used to return the names of all files that matches a pattern. This function returns an error ErrBadPattern when the pattern is malformed.
ErrBadPattern is defined in the filepath package as a global variable.
1var ErrBadPattern = errors.New("syntax error in pattern")
errors.New() is used to create a new error. We will discuss this in detail in the next tutorial.
ErrBadPattern is returned by the Glob function when the pattern is malformed.
Let’s write a small program to check for this error.
1package main
2
3import (
4 "errors"
5 "fmt"
6 "path/filepath"
7)
8
9func main() {
10 files, err := filepath.Glob("[")
11 if err != nil {
12 if errors.Is(err, filepath.ErrBadPattern) {
13 fmt.Println("Bad pattern error:", err)
14 return
15 }
16 fmt.Println("Generic error:", err)
17 return
18 }
19 fmt.Println("matched files", files)
20}
In the program above we search for files of pattern [ which is a malformed pattern. We check whether the error is not nil. To get more information about the error, we directly compare it to filepath.ErrBadPattern in line. no 11 using the Is function. Similar to As, the Is function works on an error chain. We will learn more about this in our next tutorial.
For the purposes of this tutorial, the Is function can be thought of as returning true if both the errors passed to it are the same.
The Is returns true in line no. 12 since the error is due to a malformed pattern. This program will print,
Bad pattern error: syntax error in pattern
The standard library uses any of the above-mentioned ways to provide more information about an error. We will use these ways in the next tutorial to create our own custom errors.
Do not ignore errors
Never ever ignore an error. Ignoring errors is inviting for trouble. Let me rewrite the example which lists the name of all files that match a pattern ignoring errors.
1package main
2
3import (
4 "fmt"
5 "path/filepath"
6)
7
8func main() {
9 files, _ := filepath.Glob("[")
10 fmt.Println("matched files", files)
11}
We already know from the previous example that the pattern is invalid. I have ignored the error returned by the Glob function by using the _ blank identifier in line no. 9. I simply print the matched files in line no. 10. This program will print,
matched files []
Since we ignored the error, the output seems as if no files have matched the pattern but actually the pattern itself is malformed. So never ignore errors.
This brings us to the end of this tutorial.
In this tutorial, we discussed how to handle errors that occur in our program and also how to inspect the errors to get more information from them. A quick recap of what we discussed in this tutorial,
- What are errors?
- Error representation
- Various ways of extracting more information from errors
- Do not ignore errors
In the next tutorial, we will create our own custom errors and also add more context to our custom errors.
Thanks for reading. I hope you liked this tutorial. Please leave your feedback and comments. Please consider sharing this tutorial on twitter and LinkedIn. Have a good day.
If you would like to advertise on this website, hire me, or if you have any other software development needs please email me at naveen[at]golangbot[dot]com.
Next tutorial - Custom Errors
