输入文件包含邻接列表，并以以下格式显示多行:
import java.io.IOException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
public class Friends
{
public class FriendsMap extends Mapper < LongWritable, Text, Text, IntWritable >
{
private Text friendsAB;
private Text friendsBA;
private IntWritable one = new IntWritable(1);
private IntWritable oneLess = new IntWritable(-999999999);
//@SuppressWarnings("null")
@Override
public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException
{
String friendsOfA[] = null; //This will be all of the friends of the user in this row
String oneRow[] = value.toString().split("\t,"); //Break the row up into users IDs
String userA = oneRow[0]; //This is the main user for this row
for (int i=1; i < oneRow.length; i++) //Create an array of the rest of the users in this row
{
friendsOfA[i-1] = oneRow[i];
}
for (int i=0; i < oneRow.length; i++) //Output the main user in pairs with all friends plus a lagre negative #
{
friendsAB.set(userA + " " + friendsOfA[i]);
context.write(friendsAB, oneLess);
System.out.println(friendsAB + " " + oneLess);
}
for (int i = 0; i < friendsOfA.length; i++) //Output each friend pair plus the number 1
{
for (int j = i + 1; j < friendsOfA.length; j++)
{
friendsAB.set(friendsOfA[i] + " " + friendsOfA[j]);
friendsBA.set(friendsOfA[j] + " " + friendsOfA[i]);
context.write(friendsAB, one);
context.write(friendsBA, one);
System.out.println(friendsAB + " " + one);
System.out.println(friendsBA + " " + one);
}
}
}
}
class FriendReducer extends Reducer < Text, IntWritable, Text, IntWritable >
{
private IntWritable result = new IntWritable();
@Override
public void reduce( Text key, Iterable < IntWritable > values, Context context) throws IOException, InterruptedException
{
int sum = 0;
for (IntWritable val : values)
{
sum += val.get();
}
if (sum > 1)
{
result.set( sum);
context.write( key, result);
}
//At this point I have all pairs of users with recomenede friends and a count of how many times they each
//friend has been recomended to a user.
//I need to sort by user and then by number of recomendations.
//Then print the user <tab> all recomendations with commas between them.
}
}
public static void main( String[] args) throws Exception
{
Configuration conf = new Configuration();
Job job = Job.getInstance( conf, "Friends");
job.setJarByClass(Friends.class);
FileInputFormat.addInputPath( job, new Path("input"));
FileOutputFormat.setOutputPath( job, new Path("output"));
job.setMapperClass( FriendsMap.class);
job.setCombinerClass( FriendReducer.class);
job.setReducerClass( FriendReducer.class);
job.setOutputKeyClass( Text.class);
job.setOutputValueClass( IntWritable.class);
System.exit( job.waitForCompletion( true) ? 0 : 1);
}
}
这是我在控制台中遇到的错误。
17/11/15 16:05:51 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 17/11/15 16:06:54 INFO Configuration.deprecation: session.id is deprecated. Instead, use dfs.metrics.session-id 17/11/15 16:06:54 INFO jvm.JvmMetrics: Initializing JVM Metrics with processName=JobTracker, sessionId= 17/11/15 16:06:54 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same. 17/11/15 16:06:55 WARN mapred.JobClient: No job jar file set. User classes may not be found. See JobConf(Class) or JobConf#setJar(String). 17/11/15 16:06:55 INFO input.FileInputFormat: Total input paths to process : 2 17/11/15 16:07:05 INFO mapred.JobClient: Running job: job_local426825952_0001 17/11/15 16:07:05 INFO mapred.LocalJobRunner: OutputCommitter set in config null 17/11/15 16:07:05 INFO mapred.LocalJobRunner: OutputCommitter is org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter 17/11/15 16:07:05 INFO mapred.LocalJobRunner: Waiting for map tasks 17/11/15 16:07:05 INFO mapred.LocalJobRunner: Starting task: attempt_local426825952_0001_m_000000_0 17/11/15 16:07:05 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/15 16:07:05 INFO util.ProcessTree: setsid exited with exit code 0 17/11/15 16:07:05 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@670217f0 17/11/15 16:07:05 INFO mapred.LocalJobRunner: Starting task: attempt_local426825952_0001_m_000001_0 17/11/15 16:07:05 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/15 16:07:05 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@1688e9ff 17/11/15 16:07:05 INFO mapred.LocalJobRunner: Map task executor complete. 17/11/15 16:07:05 WARN mapred.LocalJobRunner: job_local426825952_0001 java.lang.Exception: java.lang.RuntimeException: java.lang.NoSuchMethodException: Friends$FriendsMap.() at org.apache.hadoop.mapred.LocalJobRunner$Job.run(LocalJobRunner.java:406) Caused by: java.lang.RuntimeException: java.lang.NoSuchMethodException: Friends$FriendsMap.() at org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:131) at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:631) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:330) at org.apache.hadoop.mapred.LocalJobRunner$Job$MapTaskRunnable.run(LocalJobRunner.java:268) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) at java.util.concurrent.FutureTask.run(FutureTask.java:262) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) at java.lang.Thread.run(Thread.java:745) Caused by: java.lang.NoSuchMethodException: Friends$FriendsMap.() at java.lang.Class.getConstructor0(Class.java:2849) at java.lang.Class.getDeclaredConstructor(Class.java:2053) at org.apache.hadoop.util.ReflectionUtils.newInstance(ReflectionUtils.java:125) ... 8 more 17/11/15 16:07:06 INFO mapred.JobClient: map 0% reduce 0% 17/11/15 16:07:06 INFO mapred.JobClient: Job complete: job_local426825952_0001 17/11/15 16:07:06 INFO mapred.JobClient: Counters: 0
将类更改为静态后，这是新的错误。
17/11/16 04:28:50 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 17/11/16 04:28:52 INFO Configuration.deprecation: session.id is deprecated. Instead, use dfs.metrics.session-id 17/11/16 04:28:52 INFO jvm.JvmMetrics: Initializing JVM Metrics with processName=JobTracker, sessionId= 17/11/16 04:28:52 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same. 17/11/16 04:28:52 WARN mapred.JobClient: No job jar file set. User classes may not be found. See JobConf(Class) or JobConf#setJar(String). 17/11/16 04:28:53 INFO input.FileInputFormat: Total input paths to process : 2 17/11/16 04:28:54 INFO mapred.LocalJobRunner: OutputCommitter set in config null 17/11/16 04:28:54 INFO mapred.JobClient: Running job: job_local1593958162_0001 17/11/16 04:28:54 INFO mapred.LocalJobRunner: OutputCommitter is org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter 17/11/16 04:28:54 INFO mapred.LocalJobRunner: Waiting for map tasks 17/11/16 04:28:54 INFO mapred.LocalJobRunner: Starting task: attempt_local1593958162_0001_m_000000_0 17/11/16 04:28:54 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/16 04:28:54 INFO util.ProcessTree: setsid exited with exit code 0 17/11/16 04:28:54 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@57d51956 17/11/16 04:28:54 INFO mapred.MapTask: Processing split: file:/home/cloudera/workspace/Assignment4/input/Sample4.txt:0+4106187 17/11/16 04:28:54 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer 17/11/16 04:28:54 INFO mapred.MapTask: io.sort.mb = 100 17/11/16 04:28:55 INFO mapred.MapTask: data buffer = 79691776/99614720 17/11/16 04:28:55 INFO mapred.MapTask: record buffer = 262144/327680 17/11/16 04:28:55 INFO mapred.LocalJobRunner: Starting task: attempt_local1593958162_0001_m_000001_0 17/11/16 04:28:55 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/16 04:28:55 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@774140b3 17/11/16 04:28:55 INFO mapred.MapTask: Processing split: file:/home/cloudera/workspace/Assignment4/input/Sample4.txt~:0+0 17/11/16 04:28:55 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer 17/11/16 04:28:55 INFO mapred.MapTask: io.sort.mb = 100 17/11/16 04:28:55 INFO mapred.JobClient: map 0% reduce 0% 17/11/16 04:28:55 INFO mapred.MapTask: data buffer = 79691776/99614720 17/11/16 04:28:55 INFO mapred.MapTask: record buffer = 262144/327680 17/11/16 04:28:55 INFO mapred.LocalJobRunner: 17/11/16 04:28:55 INFO mapred.MapTask: Starting flush of map output 17/11/16 04:28:55 INFO mapred.Task: Task:attempt_local1593958162_0001_m_000001_0 is done. And is in the process of commiting 17/11/16 04:28:55 INFO mapred.LocalJobRunner: 17/11/16 04:28:55 INFO mapred.Task: Task 'attempt_local1593958162_0001_m_000001_0' done. 17/11/16 04:28:55 INFO mapred.LocalJobRunner: Finishing task: attempt_local1593958162_0001_m_000001_0 17/11/16 04:28:55 INFO mapred.LocalJobRunner: Map task executor complete. 17/11/16 04:28:55 WARN mapred.LocalJobRunner: job_local1593958162_0001 java.lang.Exception: java.lang.NullPointerException at org.apache.hadoop.mapred.LocalJobRunner$Job.run(LocalJobRunner.java:406) Caused by: java.lang.NullPointerException at Friends$FriendsMap.map(Friends.java:36) at Friends$FriendsMap.map(Friends.java:1) at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:140) at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:672) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:330) at org.apache.hadoop.mapred.LocalJobRunner$Job$MapTaskRunnable.run(LocalJobRunner.java:268) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) at java.util.concurrent.FutureTask.run(FutureTask.java:262) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) at java.lang.Thread.run(Thread.java:745) 17/11/16 04:28:56 INFO mapred.JobClient: Job complete: job_local1593958162_0001 17/11/16 04:28:56 INFO mapred.JobClient: Counters: 16 17/11/16 04:28:56 INFO mapred.JobClient: File System Counters 17/11/16 04:28:56 INFO mapred.JobClient: FILE: Number of bytes read=4674 17/11/16 04:28:56 INFO mapred.JobClient: FILE: Number of bytes written=139416 17/11/16 04:28:56 INFO mapred.JobClient: FILE: Number of read operations=0 17/11/16 04:28:56 INFO mapred.JobClient:
FILE: Number of large read operations=0 17/11/16 04:28:56 INFO mapred.JobClient: FILE: Number of write operations=0 17/11/16 04:28:56 INFO mapred.JobClient: Map-Reduce Framework 17/11/16 04:28:56 INFO mapred.JobClient: Map input records=0 17/11/16 04:28:56 INFO mapred.JobClient: Map output records=0 17/11/16 04:28:56 INFO mapred.JobClient: Map output bytes=0 17/11/16 04:28:56 INFO mapred.JobClient: Input split bytes=125 17/11/16 04:28:56 INFO mapred.JobClient: Combine input records=0 17/11/16 04:28:56 INFO mapred.JobClient: Combine output records=0 17/11/16 04:28:56 INFO mapred.JobClient: Spilled Records=0 17/11/16 04:28:56 INFO mapred.JobClient: CPU time spent (ms)=0 17/11/16 04:28:56 INFO mapred.JobClient: Physical memory (bytes) snapshot=0 17/11/16 04:28:56 INFO mapred.JobClient: Virtual memory (bytes) snapshot=0 17/11/16 04:28:56 INFO mapred.JobClient: Total committed heap usage (bytes)=363696128
我认为这是最麻烦的部分。
Screen Shot of some errors
这是更新的代码。
public static class FriendsMap extends Mapper < LongWritable, Text, Text, IntWritable >
{
//@SuppressWarnings("null")
@Override
public void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException
{
String friendsOfA[]; //This will be all of the friends of the user in this row
friendsOfA = new String[] {};
String friendsAB = "1"; //This will be used to create pairs of users
String friendsBA = "2"; //This will be used to create pairs of users
Text pairA;
Text pairB;
IntWritable one = new IntWritable(1); //1 if they are not an existing pair here
IntWritable oneLess = new IntWritable(-999999999); // if they are an existing pair
String oneRow[] = value.toString().split("\t,"); //Break the row up into users IDs
Text userA = new Text(oneRow[0]); //This is the main user for this row
for (int i=1; i < oneRow.length; i++) //Create an array of the rest of the users in this row
{
friendsOfA[i-1] = oneRow[i];
}
for (int i=0; i < oneRow.length; i++) //Output the main user in pairs with all friends plus a large negative #
{ //We do not want to recommend them as friends because they are friends
Text FOA = new Text (friendsOfA[i]);
friendsAB = (userA + " " + FOA);
Text pair = new Text (friendsAB);
context.write(pair, oneLess);
System.out.println(pair + " " + oneLess);
}
for (int i = 0; i < friendsOfA.length; i++) //Output each friend pair plus the number 1
{ //We want to recommend them as potential friends
for (int j = i + 1; j < friendsOfA.length; j++)
{
Text FOA = new Text (friendsOfA[i]);
Text FOB = new Text (friendsOfA[j]);
friendsAB = (FOA + " " + FOB);
friendsBA = (FOB + " " + FOA);
pairA = new Text (friendsAB);
pairB = new Text (friendsBA);
context.write(pairA, one);
context.write(pairB, one);
System.out.println(pairA + " " + one);
System.out.println(pairB + " " + one);
}
}
}
}
这是新的错误集。
17/11/16 11:59:25 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 17/11/16 11:59:27 INFO Configuration.deprecation: session.id is deprecated. Instead, use dfs.metrics.session-id 17/11/16 11:59:27 INFO jvm.JvmMetrics: Initializing JVM Metrics with processName=JobTracker, sessionId= 17/11/16 11:59:27 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same. 17/11/16 11:59:27 WARN mapred.JobClient: No job jar file set. User classes may not be found. See JobConf(Class) or JobConf#setJar(String). 17/11/16 11:59:27 INFO input.FileInputFormat: Total input paths to process : 2 17/11/16 11:59:29 INFO mapred.JobClient: Running job: job_local1899187381_0001 17/11/16 11:59:29 INFO mapred.LocalJobRunner: OutputCommitter set in config null 17/11/16 11:59:29 INFO mapred.LocalJobRunner: OutputCommitter is org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter 17/11/16 11:59:29 INFO mapred.LocalJobRunner: Waiting for map tasks 17/11/16 11:59:29 INFO mapred.LocalJobRunner: Starting task: attempt_local1899187381_0001_m_000000_0 17/11/16 11:59:29 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/16 11:59:29 INFO util.ProcessTree: setsid exited with exit code 0 17/11/16 11:59:29 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@4f94aaa1 17/11/16 11:59:29 INFO mapred.MapTask: Processing split: file:/home/cloudera/workspace/Assignment4/input/Sample4.txt:0+4106187 17/11/16 11:59:29 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer 17/11/16 11:59:29 INFO mapred.MapTask: io.sort.mb = 100 17/11/16 11:59:29 INFO mapred.MapTask: data buffer = 79691776/99614720 17/11/16 11:59:29 INFO mapred.MapTask: record buffer = 262144/327680 17/11/16 11:59:29 INFO mapred.LocalJobRunner: Starting task: attempt_local1899187381_0001_m_000001_0 17/11/16 11:59:29 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead 17/11/16 11:59:29 INFO mapred.Task: Using ResourceCalculatorPlugin : org.apache.hadoop.util.LinuxResourceCalculatorPlugin@622ecc38 17/11/16 11:59:29 INFO mapred.MapTask: Processing split: file:/home/cloudera/workspace/Assignment4/input/Sample4.txt~:0+0 17/11/16 11:59:29 INFO mapred.MapTask: Map output collector class = org.apache.hadoop.mapred.MapTask$MapOutputBuffer 17/11/16 11:59:29 INFO mapred.MapTask: io.sort.mb = 100 17/11/16 11:59:30 INFO mapred.JobClient: map 0% reduce 0% 17/11/16 11:59:30 INFO mapred.MapTask: data buffer = 79691776/99614720 17/11/16 11:59:30 INFO mapred.MapTask: record buffer = 262144/327680 17/11/16 11:59:30 INFO mapred.LocalJobRunner: 17/11/16 11:59:30 INFO mapred.MapTask: Starting flush of map output 17/11/16 11:59:30 INFO mapred.Task: Task:attempt_local1899187381_0001_m_000001_0 is done. And is in the process of commiting 17/11/16 11:59:30 INFO mapred.LocalJobRunner: 17/11/16 11:59:30 INFO mapred.Task: Task 'attempt_local1899187381_0001_m_000001_0' done. 17/11/16 11:59:30 INFO mapred.LocalJobRunner: Finishing task: attempt_local1899187381_0001_m_000001_0 17/11/16 11:59:30 INFO mapred.LocalJobRunner: Map task executor complete. 17/11/16 11:59:30 WARN mapred.LocalJobRunner: job_local1899187381_0001 java.lang.Exception: java.lang.ArrayIndexOutOfBoundsException: 0 at org.apache.hadoop.mapred.LocalJobRunner$Job.run(LocalJobRunner.java:406) Caused by: java.lang.ArrayIndexOutOfBoundsException: 0 at Friends$FriendsMap.map(Friends.java:41) at Friends$FriendsMap.map(Friends.java:1) at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:140) at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:672) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:330) at org.apache.hadoop.mapred.LocalJobRunner$Job$MapTaskRunnable.run(LocalJobRunner.java:268) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471) at java.util.concurrent.FutureTask.run(FutureTask.java:262) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) at java.lang.Thread.run(Thread.java:745) 17/11/16 11:59:31 INFO mapred.JobClient: Job complete: job_local1899187381_0001 17/11/16 11:59:31 INFO mapred.JobClient: Counters: 16 17/11/16 11:59:31 INFO mapred.JobClient: File System Counters 17/11/16 11:59:31 INFO mapred.JobClient: FILE: Number of bytes read=4674 17/11/16 11:59:31 INFO mapred.JobClient: FILE: Number of bytes written=139416 17/11/16 11:59:31 INFO mapred.JobClient: FILE: Number of read operations=0 17/11/16 11:59:31 INFO mapred.JobClient:
FILE: Number of large read operations=0 17/11/16 11:59:31 INFO mapred.JobClient: FILE: Number of write operations=0 17/11/16 11:59:31 INFO mapred.JobClient: Map-Reduce Framework 17/11/16 11:59:31 INFO mapred.JobClient: Map input records=0 17/11/16 11:59:31 INFO mapred.JobClient: Map output records=0 17/11/16 11:59:31 INFO mapred.JobClient: Map output bytes=0 17/11/16 11:59:31 INFO mapred.JobClient: Input split bytes=125 17/11/16 11:59:31 INFO mapred.JobClient: Combine input records=0 17/11/16 11:59:31 INFO mapred.JobClient: Combine output records=0 17/11/16 11:59:31 INFO mapred.JobClient: Spilled Records=0 17/11/16 11:59:31 INFO mapred.JobClient: CPU time spent (ms)=0 17/11/16 11:59:31 INFO mapred.JobClient: Physical memory (bytes) snapshot=0 17/11/16 11:59:31 INFO mapred.JobClient: Virtual memory (bytes) snapshot=0 17/11/16 11:59:31 INFO mapred.JobClient: Total committed heap usage (bytes)=363618304
最佳答案
您已将这些类视为内部类，这可能会引起问题。内部类只能存在于enclosing class的实例中。
将它们更改为静态类可能更容易。
public class Friends {
public static class FriendsMap extends Mapper <...> {}
public static class FriendReducer extends Reducer <...> {}
public static void main( String[] args) throws Exception {
Configuration conf = new Configuration();
Job job = Job.getInstance(conf, "Friends");
job.setJarByClass(Friends.class);
FileInputFormat.addInputPath(job, new Path("input"));
FileOutputFormat.setOutputPath(job, new Path("output"));
job.setMapperClass(FriendsMap.class);
job.setCombinerClass(FriendReducer.class);
job.setReducerClass(FriendReducer.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(IntWritable.class);
System.exit( job.waitForCompletion( true) ? 0 : 1);
}
}
关于java - Hadoop中的MapReduce程序，实现了一个简单的 “People You Might Know”，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/47319546/