parser.cs

csv解析类

using System;
using System.Collections.Generic;
using System.Text;

namespace QiHe.CodeLib.Csv
{
	public partial class Parser
	{
		public CsvData ParseCsvData(ParserInput<Char> input, out bool success)
		{
			this.SetInput(input);
			CsvData csvData = ParseCsvData(out success);
			return csvData;
		}

		private CsvData ParseCsvData(out bool success)
		{
			CsvData csvData = new CsvData();
			int start_position = position;

			while (true)
			{
				Record record = ParseRecord(out success);
				if (success)
				{
					csvData.Records.Add(record);
				}
				else { break; }
			}
			success = true;

			ParseEndOfFile(out success);
			if (!success)
			{
				Error("Failed to parse EndOfFile of CsvData.");
				position = start_position;
			}

			return csvData;
		}

		private Record ParseRecord(out bool success)
		{
			Record record = new Record();
			int start_position = position;

			while (true)
			{
				int seq_start_position = position;
				string str = ParseField(out success);
				if (success)
				{
					record.Fields.Add(str);
				}
				else { break; }

				while (true)
				{
					while (true)
					{
						ParseSeparator(out success);
						if (!success) { break; }

						str = ParseField(out success);
						if (success)
						{
							record.Fields.Add(str);
						}
						else { break; }

						break;
					}
					if (!success) { break; }
				}
				success = true;

				break;
			}
			if (!success)
			{
				Error("Failed to parse Fields of Record.");
				position = start_position;
				return record;
			}

			ParseEnfOfLine(out success);
			if (!success)
			{
				Error("Failed to parse EnfOfLine of Record.");
				position = start_position;
			}

			return record;
		}

		private string ParseField(out bool success)
		{
			string str = null;
			int start_position = position;

			ParseSpacing(out success);

			while (true)
			{
				str = ParseQuotedText(out success);
				if (success) { ClearError(); break; }

				str = ParseUnQuotedText(out success);
				if (success) { ClearError(); break; }

				break;
			}
			if (!success)
			{
				Error("Failed to parse (QuotedText / UnQuotedText) of Field.");
				position = start_position;
			}

			return str;
		}

		private string ParseUnQuotedText(out bool success)
		{
			string str = null;

			List<char> chars = new List<char>();
			while (true)
			{
				char ch = MatchTerminalSet(",\"\r\n", true, out success);
				if (success)
				{
					chars.Add(ch);
				}
				else { break; }
			}
			success = true;
			if (success) { if(chars.Count==0) return null; else return new string(chars.ToArray()); }
			return str;
		}

		private string ParseQuotedText(out bool success)
		{
			string str = null;
			int start_position = position;

			MatchTerminal('"', out success);
			if (!success)
			{
				Error("Failed to parse '\\\"' of QuotedText.");
				position = start_position;
				return str;
			}

			List<char> chars = new List<char>();
			while (true)
			{
				char ch;
				while (true)
				{
					ch = MatchTerminalSet("\"\r\n\\", true, out success);
					if (success) { ClearError(); break; }

					ch = ParseEscapeSequence(out success);
					if (success) { ClearError(); break; }

					break;
				}
				if (success)
				{
					chars.Add(ch);
				}
				else { break; }
			}
			success = true;

			MatchTerminal('"', out success);
			if (success) { return new string(chars.ToArray()); }
			else
			{
				Error("Failed to parse '\\\"' of QuotedText.");
				position = start_position;
			}

			return str;
		}

		private char ParseEscapeSequence(out bool success)
		{
			char ch = default(char);

			MatchTerminalString("\\\\", out success);
			if (success) { return '\\'; }

			MatchTerminalString("\\\"", out success);
			if (success) { return '\"'; }

			MatchTerminalString("\\r", out success);
			if (success) { return '\r'; }

			MatchTerminalString("\\n", out success);
			if (success) { return '\n'; }

			MatchTerminalString("\\t", out success);
			if (success) { return '\t'; }

			return ch;
		}

		private char ParseSpace(out bool success)
		{
			char ch = default(char);

			ch = MatchTerminal(' ', out success);
			if (success) { ClearError(); return ch; }

			ch = MatchTerminal('\t', out success);
			if (success) { ClearError(); return ch; }

			return ch;
		}

		private void ParseSpacing(out bool success)
		{
			while (true)
			{
				ParseSpace(out success);
				if (!success) { break; }
			}
			success = true;
		}

		private void ParseSeparator(out bool success)
		{
			MatchTerminal(',', out success);
			if (!success)
			{
				Error("Failed to parse ',' in Separator.");
			}
		}

		private void ParseEnfOfLine(out bool success)
		{
			MatchTerminalString("\r\n", out success);
			if (success) { return; }

			MatchTerminal('\r', out success);
			if (success) { return; }

			MatchTerminal('\n', out success);
			if (success) { return; }
		}

		private void ParseEndOfFile(out bool success)
		{
			success = !Input.HasInput(position);
			if (!success)
			{
				Error("Failed to parse EndOfInput in EndOfFile.");
			}
		}

	}
}