Skip to content

Add CustomQuoteCharacter for reading csv files #689

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
appel1 merged 1 commit into from
Oct 15, 2024
Merged

Add CustomQuoteCharacter for reading csv files #689

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
124 changes: 99 additions & 25 deletions src/ExcelDataReader.Tests/ExcelCsvReaderTest.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
namespace ExcelDataReader.Tests;

/// <summary>
/// Most CSV test data came from csv-spectrum: https://github.com/maxogden/csv-spectrum.
/// Most CSV test data came from csv-spectrum: https://github.com/maxogden/csv-spectrum.
/// </summary>
public class ExcelCsvReaderTest
{
Expand Down Expand Up @@ -187,7 +187,7 @@ public void CsvWrongEncoding()
};

using (ExcelReaderFactory.CreateCsvReader(Configuration.GetTestWorkbook("csv\\cp1252.csv"), configuration))
{
{
}
});
}
Expand Down Expand Up @@ -354,20 +354,18 @@ public void CsvRowCountAnalyzeRowsThrows()
public void GitIssue578()
{
var stream = Configuration.GetTestWorkbook(@"csv\Test_git_issue578.csv");
using (IExcelDataReader excelReader = ExcelReaderFactory.CreateCsvReader(stream, new ExcelReaderConfiguration()))
{
excelReader.Read();
var values = new object[excelReader.FieldCount];
excelReader.GetValues(values);
var values2 = new object[excelReader.FieldCount + 1];
excelReader.GetValues(values2);
var values3 = new object[excelReader.FieldCount - 1];
excelReader.GetValues(values3);

Assert.That(values, Is.EqualTo(new object[] { "1", "2", "3", "4", "5" }));
Assert.That(values2, Is.EqualTo(new object[] { "1", "2", "3", "4", "5", null }));
Assert.That(values3, Is.EqualTo(new object[] { "1", "2", "3", "4" }));
}
using IExcelDataReader excelReader = ExcelReaderFactory.CreateCsvReader(stream, new ExcelReaderConfiguration());
excelReader.Read();
var values = new object[excelReader.FieldCount];
excelReader.GetValues(values);
var values2 = new object[excelReader.FieldCount + 1];
excelReader.GetValues(values2);
var values3 = new object[excelReader.FieldCount - 1];
excelReader.GetValues(values3);

Assert.That(values, Is.EqualTo(new object[] { "1", "2", "3", "4", "5" }));
Assert.That(values2, Is.EqualTo(new object[] { "1", "2", "3", "4", "5", null }));
Assert.That(values3, Is.EqualTo(new object[] { "1", "2", "3", "4" }));
}

[Test]
Expand All @@ -377,7 +375,6 @@ public void GitIssue500()
Item_Number,Pmt_Amount,Type,Voided,Note
200212812,$462.76,Check,06/06/2018,Hash#hash


""";

using var reader = ExcelReaderFactory.CreateCsvReader(new MemoryStream(Encoding.UTF8.GetBytes(data)));
Expand All @@ -400,11 +397,10 @@ public void GitIssue500_QuotedValueWithNewLine()
{
const string data = """
Item_Number,Pmt_Amount,"Type

2",Voided,Note
200212812,$462.76,Check,06/06/2018,Hash#hash


""";

using var reader = ExcelReaderFactory.CreateCsvReader(new MemoryStream(Encoding.UTF8.GetBytes(data)));
Expand All @@ -417,15 +413,15 @@ public void GitIssue500_QuotedValueWithNewLine()

Assert.Multiple(() =>
{
Assert.That(row1, Is.EqualTo(new object[]
Assert.That(row1, Is.EqualTo(new object[]
{
"Item_Number",
"Item_Number",
"Pmt_Amount", """
Type

2
""",
"Voided",
""",
"Voided",
"Note",
}));
Assert.That(row2, Is.EqualTo(new object[] { "200212812", "$462.76", "Check", "06/06/2018", "Hash#hash" }));
Expand Down Expand Up @@ -479,4 +475,82 @@ public void GitIssue642_ActiveSheet()
Assert.That(reader.ActiveSheet, Is.EqualTo(0));
Assert.That(dataSet.Tables.Count, Is.EqualTo(1));
}
}

[Test]
public void GitIssue580_ReadCsvWithoutQuoteChar()
{
#pragma warning disable SA1027 // TabsMustNotBeUsed
const string data = """
Name Currency Ty"pe Cost Cost
Test1 A"BC XX 10,143.27 "0.00
Test2 EFG YY 10,143.27 0.00
Test3 IJK ZZ" 10,143.27 0.00

""";
#pragma warning restore SA1027 // TabsMustNotBeUsed

using var reader = ExcelReaderFactory.CreateCsvReader(new MemoryStream(Encoding.UTF8.GetBytes(data)), new()
{
QuoteChar = null
});
reader.Read();
object[] row1 = new object[reader.FieldCount];
reader.GetValues(row1);
reader.Read();
object[] row2 = new object[reader.FieldCount];
reader.GetValues(row2);
reader.Read();
object[] row3 = new object[reader.FieldCount];
reader.GetValues(row3);
reader.Read();
object[] row4 = new object[reader.FieldCount];
reader.GetValues(row4);

Assert.Multiple(() =>
{
Assert.That(row1, Is.EqualTo(new object[] { "Name", "Currency", "Ty\"pe", "Cost", "Cost" }));
Assert.That(row2, Is.EqualTo(new object[] { "Test1", "A\"BC", "XX", "10,143.27", "\"0.00" }));
Assert.That(row3, Is.EqualTo(new object[] { "Test2", "EFG", "YY", "10,143.27", "0.00" }));
Assert.That(row4, Is.EqualTo(new object[] { "Test3", "IJK", "ZZ\"", "10,143.27", "0.00" }));
});
}

[Test]
public void GitIssue580_ReadCsvWithCustomQuoteChar()
{
#pragma warning disable SA1027 // TabsMustNotBeUsed
const string data = """
Name Currency Type Cost 'Cost per 1,000 Items'
Test1 ABC XX '10,143.27' 0.00
Test2 EFG YY '10,143.27' 0.00
Test3 IJK ZZ '10,143.27' 0.00

""";
#pragma warning restore SA1027 // TabsMustNotBeUsed

using var reader = ExcelReaderFactory.CreateCsvReader(new MemoryStream(Encoding.UTF8.GetBytes(data)), new()
{
QuoteChar = '\''
});
reader.Read();
object[] row1 = new object[reader.FieldCount];
reader.GetValues(row1);
reader.Read();
object[] row2 = new object[reader.FieldCount];
reader.GetValues(row2);
reader.Read();
object[] row3 = new object[reader.FieldCount];
reader.GetValues(row3);
reader.Read();
object[] row4 = new object[reader.FieldCount];
reader.GetValues(row4);

Assert.Multiple(() =>
{
Assert.That(row1, Is.EqualTo(new object[] { "Name", "Currency", "Type", "Cost", "Cost per 1,000 Items" }));
Assert.That(row2, Is.EqualTo(new object[] { "Test1", "ABC", "XX", "10,143.27", "0.00" }));
Assert.That(row3, Is.EqualTo(new object[] { "Test2", "EFG", "YY", "10,143.27", "0.00" }));
Assert.That(row4, Is.EqualTo(new object[] { "Test3", "IJK", "ZZ", "10,143.27", "0.00" }));
});
}
}
8 changes: 4 additions & 4 deletions src/ExcelDataReader/Core/CsvFormat/CsvAnalyzer.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,11 +5,11 @@ namespace ExcelDataReader.Core.CsvFormat;
internal static class CsvAnalyzer
{
/// <summary>
/// Reads completely through a CSV stream to determine encoding, separator, field count and row count.
/// Reads completely through a CSV stream to determine encoding, separator, field count and row count.
/// Uses fallbackEncoding if there is no BOM. Throws DecoderFallbackException if there are invalid characters in the stream.
/// Returns the separator whose average field count is closest to its max field count.
/// </summary>
public static void Analyze(Stream stream, char[] separators, Encoding fallbackEncoding, int analyzeInitialCsvRows, out int fieldCount, out char autodetectSeparator, out Encoding autodetectEncoding, out int bomLength, out int rowCount)
public static void Analyze(Stream stream, char[] separators, Encoding fallbackEncoding, int analyzeInitialCsvRows, char? quoteChar, out int fieldCount, out char autodetectSeparator, out Encoding autodetectEncoding, out int bomLength, out int rowCount)
{
var bufferSize = 1024;
var probeSize = 16;
Expand All @@ -29,7 +29,7 @@ public static void Analyze(Stream stream, char[] separators, Encoding fallbackEn
{
separatorInfos[i] = new SeparatorInfo
{
Buffer = new CsvParser(separators[i], autodetectEncoding)
Buffer = new CsvParser(separators[i], autodetectEncoding, quoteChar)
};
}

Expand Down Expand Up @@ -191,4 +191,4 @@ private sealed class SeparatorInfo

public CsvParser Buffer { get; set; }
}
}
}
22 changes: 14 additions & 8 deletions src/ExcelDataReader/Core/CsvFormat/CsvParser.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,10 +7,10 @@ namespace ExcelDataReader.Core.CsvFormat;
/// </summary>
internal sealed class CsvParser
{
public CsvParser(char separator, Encoding encoding)
public CsvParser(char separator, Encoding encoding, char? quoteChar = null)
{
Separator = separator;
QuoteChar = '"';
QuoteChar = quoteChar;

Decoder = encoding.GetDecoder();
Decoder.Fallback = new DecoderExceptionFallback();
Expand All @@ -34,7 +34,7 @@ private enum CsvState

private CsvState State { get; set; }

private char QuoteChar { get; }
private char? QuoteChar { get; }

private int TrailingWhitespaceCount { get; set; }

Expand Down Expand Up @@ -93,21 +93,27 @@ private void ParseChar(char c, int bytesUsed)
case CsvState.PreValue:
parsed = ReadPreValue(c, bytesUsed);
break;

case CsvState.Value:
parsed = ReadValue(c, bytesUsed);
break;

case CsvState.QuotedValue:
parsed = ReadQuotedValue(c, bytesUsed);
break;

case CsvState.QuotedValueQuote:
parsed = ReadQuotedValueQuote(c, bytesUsed);
break;

case CsvState.Separator:
parsed = ReadSeparator(c, bytesUsed);
break;

case CsvState.Linebreak:
parsed = ReadLinebreak(c, bytesUsed);
break;

default:
throw new InvalidOperationException("Unhandled parser state: " + State);
}
Expand All @@ -120,7 +126,7 @@ private bool ReadPreValue(char c, int bytesUsed)
{
return true;
}
else if (c == QuoteChar)
else if (QuoteChar.HasValue && c == QuoteChar.Value)
{
State = CsvState.QuotedValue;
return true;
Expand Down Expand Up @@ -182,7 +188,7 @@ private bool ReadValue(char c, int bytesUsed)

private bool ReadQuotedValue(char c, int bytesUsed)
{
if (c == QuoteChar)
if (QuoteChar.HasValue && c == QuoteChar.Value)
{
State = CsvState.QuotedValueQuote;
return true;
Expand All @@ -196,7 +202,7 @@ private bool ReadQuotedValue(char c, int bytesUsed)

private bool ReadQuotedValueQuote(char c, int bytesUsed)
{
if (c == QuoteChar)
if (QuoteChar.HasValue && c == QuoteChar.Value)
{
// Is escaped quote
ValueResult.Append(c);
Expand Down Expand Up @@ -244,7 +250,7 @@ private bool ReadLinebreak(char c, int bytesUsed)

private void AddValueToRow()
{
RowResult.Add(ValueResult.ToString(0, ValueResult.Length - TrailingWhitespaceCount));
RowResult.Add(ValueResult.ToString(0, ValueResult.Length - TrailingWhitespaceCount));
ValueResult = new StringBuilder();
TrailingWhitespaceCount = 0;
}
Expand All @@ -269,4 +275,4 @@ private bool IsWhitespace(char c)

return false;
}
}
}
8 changes: 5 additions & 3 deletions src/ExcelDataReader/Core/CsvFormat/CsvWorkbook.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@

namespace ExcelDataReader.Core.CsvFormat;

internal sealed class CsvWorkbook(Stream stream, Encoding encoding, char[] autodetectSeparators, int analyzeInitialCsvRows) : IWorkbook<CsvWorksheet>
internal sealed class CsvWorkbook(Stream stream, Encoding encoding, char[] autodetectSeparators, int analyzeInitialCsvRows, char? quoteChar = null) : IWorkbook<CsvWorksheet>
{
public int ResultsCount => 1;

Expand All @@ -13,17 +13,19 @@ internal sealed class CsvWorkbook(Stream stream, Encoding encoding, char[] autod

public Encoding Encoding { get; } = encoding;

public char? QuoteChar { get; } = quoteChar;

public char[] AutodetectSeparators { get; } = autodetectSeparators;

public int AnalyzeInitialCsvRows { get; } = analyzeInitialCsvRows;

public IEnumerable<CsvWorksheet> ReadWorksheets()
{
yield return new CsvWorksheet(Stream, Encoding, AutodetectSeparators, AnalyzeInitialCsvRows);
yield return new CsvWorksheet(Stream, Encoding, AutodetectSeparators, AnalyzeInitialCsvRows, QuoteChar);
}

public NumberFormatString GetNumberFormatString(int index)
{
return null;
}
}
}
13 changes: 8 additions & 5 deletions src/ExcelDataReader/Core/CsvFormat/CsvWorksheet.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,14 +4,15 @@ namespace ExcelDataReader.Core.CsvFormat;

internal sealed class CsvWorksheet : IWorksheet
{
public CsvWorksheet(Stream stream, Encoding fallbackEncoding, char[] autodetectSeparators, int analyzeInitialCsvRows)
public CsvWorksheet(Stream stream, Encoding fallbackEncoding, char[] autodetectSeparators, int analyzeInitialCsvRows, char? quoteChar = null)
{
Stream = stream;
QuoteChar = quoteChar;
Stream.Seek(0, SeekOrigin.Begin);
try
{
// Try as UTF-8 first, or use BOM if present
CsvAnalyzer.Analyze(Stream, autodetectSeparators, Encoding.UTF8, analyzeInitialCsvRows, out var fieldCount, out var separator, out var encoding, out var bomLength, out var rowCount);
CsvAnalyzer.Analyze(Stream, autodetectSeparators, Encoding.UTF8, analyzeInitialCsvRows, quoteChar, out var fieldCount, out var separator, out var encoding, out var bomLength, out var rowCount);
FieldCount = fieldCount;
AnalyzedRowCount = rowCount;
AnalyzedPartial = analyzeInitialCsvRows > 0;
Expand All @@ -24,7 +25,7 @@ public CsvWorksheet(Stream stream, Encoding fallbackEncoding, char[] autodetectS
// If cannot parse as UTF-8, try fallback encoding
Stream.Seek(0, SeekOrigin.Begin);

CsvAnalyzer.Analyze(Stream, autodetectSeparators, fallbackEncoding, analyzeInitialCsvRows, out var fieldCount, out var separator, out var encoding, out var bomLength, out var rowCount);
CsvAnalyzer.Analyze(Stream, autodetectSeparators, fallbackEncoding, analyzeInitialCsvRows, quoteChar, out var fieldCount, out var separator, out var encoding, out var bomLength, out var rowCount);
FieldCount = fieldCount;
AnalyzedRowCount = rowCount;
AnalyzedPartial = analyzeInitialCsvRows > 0;
Expand Down Expand Up @@ -67,6 +68,8 @@ public int RowCount

public Encoding Encoding { get; }

public char? QuoteChar { get; }

public char Separator { get; }

public Column[] ColumnWidths => null;
Expand All @@ -82,7 +85,7 @@ public IEnumerable<Row> ReadRows()
var bufferSize = 1024;
var buffer = new byte[bufferSize];
var rowIndex = 0;
var csv = new CsvParser(Separator, Encoding);
var csv = new CsvParser(Separator, Encoding, QuoteChar);
var skipBomBytes = BomLength;

Stream.Seek(0, SeekOrigin.Begin);
Expand Down Expand Up @@ -124,4 +127,4 @@ private static IEnumerable<Row> GetReaderRows(int rowIndex, List<List<string>> r
rowIndex++;
}
}
}
}
Loading