ByteArrayOutputStream转换为字符串数组

Question

I'm writing an application which has a method that will download a text file from my server. This text file will contain ~1,000 proxy IP's. The download will happen every 10 minutes. I need to find the most efficient way of doing this.

Currently I have a method in a class called Connection which will return the bytes of whatever I want to retrieve. So if I make a connection to the server for the text file using such method, I will get it returned in bytes. My other method will create a very long string from these bytes. After, I split the long string into an array using System.LineSeparator. Here is the code:

 public static void fetchProxies(String url) {
    Connection c = new Connection();
    List<Proxy> tempProxy = new ArrayList<Proxy>();
    ByteArrayOutputStream baos = 
            c.requestBytes(url);  
    String line = new String(baos.toByteArray()); 

    String[] split = line.split(System.lineSeparator());
    //more code to come but the above works fine.

}

This currently works but I know that it isn't the most efficient way. I

Answer 1

The most efficient and logical way would be to create a BufferedReader wrapping an InputStreamReader wrapping the InputStream of the URL connection. You would the use readLine() on the BufferedReader until it returns null, and append each line read to the list of IP addresses:

List<String> ipList = new ArrayList<>();
try (BufferedReader reader = new BufferedReader(new InputStreamReader(connection.getInputStream(), theAppropriateEncoding))) {
    String line;
    while ((line = reader.readLine()) != null) {
        ipList.add(line);
    }
}

Note that this probably won't change much in the performance of the method, though, because most of the time is being spend in waiting fof bytes coming from the remote host, which is considerably slower than building and splitting a String in memory.

Answer 2

split method from String isn't the fastest way to separate all the IP's. There ara other libraries to achive this in an more optimized way. Read this: http://demeranville.com/battle-of-the-tokenizers-delimited-text-parser-performance/

There is a very nice time comparision about 7 different ways to split a String.

For example a the Splitter class from Guava library returns an Iterable, and with Guava you could also convert the results as List:

import com.google.common.base.Splitter;
...
public static void fetchProxies(String url) {
Connection c = new Connection();
List<Proxy> tempProxy = new ArrayList<Proxy>();
ByteArrayOutputStream baos = 
        c.requestBytes(url);  
String line = new String(baos.toByteArray()); 

Iterator<Element> myIterator = 
    Splitter.on(System.getProperty("line.separator")).split(line);
List<Element> myList = Lists.newArrayList(myIterator);

// do somethjing with the List...