/// <summary>
/// Reads the lines from the reader.
/// </summary>
/// <param name="reader">The text reader to read the data from.</param>
/// <param name="options">The optional options to use when reading.</param>
public static IEnumerable<ICsvLine> Read(TextReader reader, CsvOptions options = null)
{
if (reader == null)
throw new ArgumentNullException(nameof(reader));
return ReadImpl(reader, options);
}
/// <summary>
/// Reads the lines from the stream.
/// </summary>
/// <param name="stream">The stream to read the data from.</param>
/// <param name="options">The optional options to use when reading.</param>
public static IEnumerable<ICsvLine> ReadFromStream(Stream stream, CsvOptions options = null)
{
if (stream == null)
throw new ArgumentNullException(nameof(stream));
return ReadFromStreamImpl(stream, options);
}
/// <summary>
/// Reads the lines from the csv string.
/// </summary>
/// <param name="csv">The csv string to read the data from.</param>
/// <param name="options">The optional options to use when reading.</param>
public static IEnumerable<ICsvLine> ReadFromText(string csv, CsvOptions options = null)
{
if (csv == null)
throw new ArgumentNullException(nameof(csv));
return ReadFromTextImpl(csv, options);
}
private static string[] CreateDefaultHeaders(string line, CsvOptions options)
{
var columnCount = options.Splitter.Matches(line);
var headers = new string[columnCount.Count];
for (var i = 0; i < headers.Length; i++)
headers[i] = $"Column{i + 1}";
return headers;
}
private static string[] GetHeaders(string line, CsvOptions options)
{
return SplitLine(line, options);
}
private CsvCell(object value, bool inner, CsvOptions options)
{
_options = options;
_value = value;
_inner = inner;
}
internal static string Serialize(T obj, CsvOptions options)
{
return(new CsvRow <T>(obj, options).Serialize());
}
private static string[] GetHeaders(string line, CsvOptions options)
{
return(SplitLine(line, options));
}
public static IEnumerable <T> Read <T>(TextReader reader, CsvOptions options, CsvFilter filter = null) where T : class, new()
{
return(new CsvReader <T>(reader, options, filter).Read());
}
private static IEnumerable<ICsvLine> ReadFromTextImpl(string csv, CsvOptions options)
{
using (var reader = new StringReader(csv))
{
foreach (var line in ReadImpl(reader, options))
yield return line;
}
}
private static IEnumerable<ICsvLine> ReadImpl(TextReader reader, CsvOptions options)
{
if (options == null)
options = new CsvOptions();
string line;
var index = 0;
string[] headers = null;
Dictionary<string, int> headerLookup = null;
var initalized = false;
while ((line = reader.ReadLine()) != null)
{
index++;
if (index <= options.RowsToSkip || options.SkipRow?.Invoke(line, index) == true)
continue;
if (!initalized)
{
InitalizeOptions(line, options);
var skipInitialLine = options.HeaderMode == HeaderMode.HeaderPresent;
headers = skipInitialLine ? GetHeaders(line, options) : CreateDefaultHeaders(line, options);
headerLookup = headers.Select((h, idx) => Tuple.Create(h, idx)).ToDictionary(h => h.Item1, h => h.Item2, options.Comparer);
initalized = true;
if (skipInitialLine)
continue;
}
yield return new ReadLine(headers, headerLookup, index, line, options);
}
}
private static IEnumerable <ICsvLine> ReadImpl(TextReader reader, CsvOptions options)
{
if (options == null)
{
options = new CsvOptions();
}
string line;
var index = 0;
string[] headers = null;
Dictionary <string, int> headerLookup = null;
var initalized = false;
while ((line = reader.ReadLine()) != null)
{
index++;
if (index <= options.RowsToSkip || options.SkipRow?.Invoke(line, index) == true)
{
continue;
}
if (!initalized)
{
InitializeOptions(line, options);
var skipInitialLine = options.HeaderMode == HeaderMode.HeaderPresent;
headers = skipInitialLine ? GetHeaders(line, options) : CreateDefaultHeaders(line, options);
try
{
headerLookup = headers.Select((h, idx) => Tuple.Create(h, idx)).ToDictionary(h => h.Item1, h => h.Item2, options.Comparer);
}
catch (ArgumentException)
{
throw new InvalidOperationException("Duplicate headers detected in HeaderPresent mode. If you don't have a header you can set the HeaderMode to HeaderAbsent.");
}
var aliases = options.Aliases;
if (aliases != null)
{
// NOTE: For each group we need at most 1 match (i.e. SingleOrDefault)
foreach (var aliasGroup in aliases)
{
var groupIndex = -1;
foreach (var alias in aliasGroup)
{
if (headerLookup.TryGetValue(alias, out var aliasIndex))
{
if (groupIndex != -1)
{
throw new InvalidOperationException("Found multiple matches within alias group: " + string.Join(";", aliasGroup));
}
groupIndex = aliasIndex;
}
}
if (groupIndex != -1)
{
foreach (var alias in aliasGroup)
{
headerLookup[alias] = groupIndex;
}
}
}
}
initalized = true;
if (skipInitialLine)
{
continue;
}
}
var record = new ReadLine(headers, headerLookup, index, line, options);
if (options.AllowNewLineInEnclosedFieldValues)
{
while (record.RawSplitLine.Any(f => IsUnterminatedQuotedValue(f, options)))
{
var nextLine = reader.ReadLine();
if (nextLine == null)
{
break;
}
line += options.NewLine + nextLine;
record = new ReadLine(headers, headerLookup, index, line, options);
}
}
yield return(record);
}
}
private static IEnumerable<ICsvLine> ReadFromStreamImpl(Stream stream, CsvOptions options)
{
using (var reader = new StreamReader(stream))
{
foreach (var line in ReadImpl(reader, options))
yield return line;
}
}
internal static T Deserialize(string row, PropertyInfo[] headers, CsvOptions options)
{
return(new CsvRow <T>(options).Deserialize(row, headers));
}
internal static PropertyInfo[] DeserializeHeaders(string[] headerNames, CsvOptions options)
{
return(new CsvRow <T>(options).DeserializeHeaders(headerNames));
}
internal static PropertyInfo[] DeserializeHeaders(string row, CsvOptions options)
{
return(new CsvRow <T>(options).DeserializeHeader(row));
}
private static void InitalizeOptions(string line, CsvOptions options)
{
if (options.Separator == '\0')
options.Separator = AutoDetectSeparator(line);
Regex splitter;
lock (syncRoot)
{
if (!splitterCache.TryGetValue(options.Separator, out splitter))
splitterCache[options.Separator] = splitter = new Regex(string.Format(@"(?>(?(IQ)(?(ESC).(?<-ESC>)|\\(?<ESC>))|(?!))|(?(IQ)\k<QUOTE>(?<-IQ>)|(?<QUOTE>"")(?<IQ>))|(?(IQ).|[^{0}]))+|^(?={0})|(?<={0})(?={0})|(?<={0})$", Regex.Escape(options.Separator.ToString())), (RegexOptions)8);
}
options.Splitter = splitter;
}
private static string[] SplitLine(string line, CsvOptions options)
{
var matches = options.Splitter.Matches(line);
var parts = new string[matches.Count];
for (var i = 0; i < matches.Count; i++)
{
var str = matches[i].Value;
if (options.TrimData)
str = str.Trim();
if (str.StartsWith("\"") && str.EndsWith("\""))
str = str.Substring(1, str.Length - 2);
parts[i] = str;
}
return parts;
}
private static IEnumerable <ICsvLine> ReadImpl(TextReader reader, CsvOptions options)
{
if (options == null)
{
options = new CsvOptions();
}
string line;
var index = 0;
string[] headers = null;
Dictionary <string, int> headerLookup = null;
var initalized = false;
while ((line = reader.ReadLine()) != null)
{
index++;
if (index <= options.RowsToSkip || options.SkipRow?.Invoke(line, index) == true)
{
continue;
}
if (!initalized)
{
InitializeOptions(line, options);
var skipInitialLine = options.HeaderMode == HeaderMode.HeaderPresent;
headers = skipInitialLine ? GetHeaders(line, options) : CreateDefaultHeaders(line, options);
headerLookup = headers.Select((h, idx) => Tuple.Create(h, idx)).ToDictionary(h => h.Item1, h => h.Item2, options.Comparer);
var aliases = options.Aliases;
if (aliases != null)
{
// NOTE: For each group we need at most 1 match (i.e. SingleOrDefault)
foreach (var aliasGroup in aliases)
{
var groupIndex = -1;
foreach (var alias in aliasGroup)
{
if (headerLookup.TryGetValue(alias, out var aliasIndex))
{
if (groupIndex != -1)
{
throw new InvalidOperationException("Found multiple matches within alias group: " + string.Join(";", aliasGroup));
}
groupIndex = aliasIndex;
}
}
if (groupIndex != -1)
{
foreach (var alias in aliasGroup)
{
headerLookup[alias] = groupIndex;
}
}
}
}
initalized = true;
if (skipInitialLine)
{
continue;
}
}
// TODO: #11 might need to read another line if this one isn't ended yet (i.e. open quoted value)
yield return(new ReadLine(headers, headerLookup, index, line, options));
}
}
public ReadLine(string[] headers, Dictionary<string, int> headerLookup, int index, string raw, CsvOptions options)
{
this.headerLookup = headerLookup;
this.options = options;
Headers = headers;
Raw = raw;
Index = index;
}
public ReadLine(string[] headers, Dictionary <string, int> headerLookup, int index, string raw, CsvOptions options)
{
this.headerLookup = headerLookup;
this.options = options;
Headers = headers;
Raw = raw;
Index = index;
}
internal static string Serialize(object value, CsvOptions options)
{
return(Serialize(value, false, options));
}
public static IEnumerable <T> Read <T>(TextReader reader, CsvFilter filter = null) where T : class, new()
{
return(Read <T>(reader, CsvOptions.CreateDefault(), filter));
}
private CsvRow(T obj, CsvOptions options)
{
_options = options;
Obj = obj;
}
private CsvCell(CsvOptions options)
{
_options = options;
_inner = false;
}
internal static string SerializeHeader(CsvOptions options)
{
return(new CsvRow <T>(options).SerializeHeader());
}
private static string Serialize(object value, bool inner, CsvOptions options)
{
return(new CsvCell(value, inner, options).Serialize());
}
private CsvRow(CsvOptions options)
{
_options = options;
Obj = new T(); //Not null for deserialization
}
