CSV Parser (comma delimited values) written in Java and works with EXCEL exported files

ParserForCSV.java

//-------------------------------------------------------------------------------
// This file is made available under the Creative Commons CC0 1.0 Universal Public Domain Dedication.
// The person who associated a work with this deed has dedicated the work to the public domain by
// waiving all of his or her rights to the work worldwide under copyright law, including all related
// and neighboring rights, to the extent allowed by law. You can copy, modify, distribute and perform 
// the work, even for commercial purposes, all without asking permission.

import java.io.IOException;
import java.io.Reader;
import java.io.StringReader;
import java.util.ArrayList;
import java.util.List;

//--------------------------------------------------------------------------------
// Parser for files that contain comma separated values.
// This is designed to handle EXCEL exported files.  For example
/*  ====  SAMPLE BEGIN ====
	"Name", "Age", Description
	John Henckel,52,This is a string of text
	Carin Pie,28,"This
	text is broken
	over \"several\" lines."
	Phen "Bob" Weirsby,,missing age
	====  SAMPLE END ==== */
// Notice that newline is a record delimiter, except within quotes. Also within quotes
// you can have escape sequences \n \" \\ \t. Other escape sequences are not changed.
// The " character is allowed in a non-quoted field if it is not the first character.
// Fields not in quotes are trimmed for whitespace.


public class ParserForCSV {

	//--------------------------------------------------------------------------------
	// Main function.  Read entire file into list of rows.
	// messages - Any warning messages are appended here.
	
	public static List<String[]> parse(Reader file, List<String> messages) throws IOException {
		List<String[]> result = new ArrayList<String[]>();
		int n = 10;
		for(int i = 0;; ++i) {
			String[] row = parseRecord(file, messages, i, n);
			if (row == null) break;
			result.add(row);
			if (row.length > n) n = Math.min(row.length, 2*n);
		}
		return result;
	}

	//--------------------------------------------------------------------------------
	// This reads a record from a CSV file into an array of strings.
	// Returns null for EOF.
	// The hint is the expected number of fields per record.

	public static String[] parseRecord(Reader file, List<String> messages, int rowNumber, int hint) throws IOException {
		StringBuilder sb = new StringBuilder();
		List<String> row = new ArrayList<String>(Math.min(Math.max(10, hint), 1000));
		for(int i = 0;;) {
			int c = parseField(file, sb, messages, rowNumber, i);
			if (i==0 && sb.length() == 0) {
				if (c == -1)
					return null;		// end of the file
				if (c == '\n')
					continue;        // ignore blank lines in the middle of the file
			}
			row.add(sb.toString());
			if (c != ',') break;
			++i;
		}
		return row.toArray(new String[row.size()]);
	}

	
	//--------------------------------------------------------------------------------
	// This reads a single FIELD item from a CSV file into a string buffer
	// Returns the last character read, which is comma, newline, or EOF (-1).
	// Append any warnings to messages.
	
	public static int parseField(Reader file, StringBuilder result, List<String> messages, int row, int col) throws IOException {
		String dump = null;
		result.setLength(0);
		int c = file.read();
		int numChars = 0;      // number of non-white chars that are added to the result
		for (;;) {
			if (c == '\n' || c == ',' || c == -1) {
				if (dump != null)
					messages.add("WARN: Ignoring data in record "+row+" field "+(col+1)+": " + dump);
				if (numChars > 0) {
					// trim trailing whitespace for non-quoted fields
					while (numChars > 1 && Character.isWhitespace(result.charAt(numChars - 1))) --numChars;
					result.setLength(numChars);
				}
				return c;
			}
			if (numChars == 0 && c == '"') {
				result.setLength(0);    // Ignore whitespace in front of quotes
				for (;;) {
					c = file.read();
					if (c == '"' || c == -1) {
						if (c == -1) {
							messages.add("WARN: Unexpected EOF in record "+row+" field "+(col+1));
							return c;
						}
						break;
					}
					if (c == '\\') {
						c = file.read();
						// Interpret \n, \t, \\, and \" in the strings
						if (c == 'n') c = '\n';
						else if (c == 't') c = '\t';
						else if (c != '"' && c != '\\') result.append('\\');
					}
					result.append((char)c);
				}
				// The value -1 means we finished parsing the quotes and now we're looking for a comma or newline.
				numChars = -1;  
			}
			else {
				if (numChars == -1) {
					if (dump != null) 
						dump += (char)c;
					else if (!Character.isWhitespace(c)) 
						dump = "" + (char) c;
				} else {
					if (numChars > 0 || !Character.isWhitespace(c)) {
						result.append((char)c);
						++numChars;
					}
				}
			}
			c = file.read();			
		}
	}

	
	public static void TestCase() {
		String data =
				"\"Name\", \"Age\", Description\n" + 
				"John Henckel,52,This is a string of text\n" + 
				"Thomas Moore,\"52\" years old\n" + 
				"Carin Pie,28,\"This\n" + 
				"text is broken\n" + 
				"over \\\"several\\\" lines.\"\n" + 
				"Phen \"Bob\" Weirsby,,\"missing age\n";
		List<String> messages = new ArrayList<>();
		List<String[]> result;
		try {
			result = parse(new StringReader(data), messages);
		} catch (Exception e) {
			e.printStackTrace();
			return;
		}
		int i = 0;
		System.out.println("---- result data -----");
		for (String[] row: result) {
			System.out.println("BEGIN ROW " + i++);
			for (String item: row) System.out.println("   DATA: \""+item+"\"");
		}
		System.out.println("---- parser messages -----");
		for (String m: messages) {
			System.out.println(m);
		}
		
	}
}