Reading java file with escape characters for newline

问题

I have a Unicode file that needs to be exported to database(Vertica). The column delimiter is CTRL+B, record delimiter is newline(\n). Whenever there is a newline within a column value, CTRL+A is used as escape character.

When I use BufferedReader.readLine() to read this file, the records with ID's 2 and 4, are read as two records. Whereas I want to read them as a single whole record as given in output.

Here is the example input file. | stands for CTRL+B and ^ stands for CTRL+A.

Input
ID|Name|Job Desc
----------------
1|xxxx|SO Job
2|YYYY|SO Careers^
Job
3|RRRRR|SO
4|ZZZZ^
 ZZ|SO Job
5|AAAA|YU

Output:
ID|Name|Job Desc
----------------
1|xxxx|SO Job
2|YYYY|SO Careers Job
3|RRRRR|SO
4|ZZZZ ZZ|SO Job
5|AAAA|YU

The file is huge, so I cant use StringEscapeUtils. Any suggestions on this?

回答1:

You can use a Scanner with a custom delimeter. The delimeter I use is set to match \n but not \u0001\n (where \u0001 represents CTRL+A):

try {
    PrintWriter writer = new PrintWriter("dboutput.txt");
    Scanner sc = new Scanner(new File("dbinput.txt"));
    sc.useDelimiter(Pattern.compile("^(?!.*(\\u0001\\n)).*\\n$"));
    while (sc.hasNext()) {
        writer.println(sc.next());
    }
    scanner.close();
    writer.close();
} catch (FileNotFoundException e) {
   e.printStackTrace();
} 

回答2:

Tim is partially right in his answer. But, it still does not resolve newlines escaped by CTRL+A.

Here is my solution for that(guided by Tim answer)

File f = new File("C:\\Users\\SV7104\\Desktop\\sampletest.txt");
Scanner sc = new Scanner(f).useDelimiter(Pattern.compile("\\s*\\u0002\\n\\s*"));
            while (sc.hasNext()) {
                System.out.print(1);
                System.out.println(sc.next().toString().replaceAll("\\u0001\\n", " "));

            }

If there's some other efficient method, I am curious to know about that too.

来源：https://stackoverflow.com/questions/29908236/reading-java-file-with-escape-characters-for-newline