Changeset 13974 for stable/HeuristicLab.Tests/HeuristicLab.Problems.Instances.DataAnalysis-3.3

Timestamp:

07/02/16 08:15:07 (8 years ago)

Author:

gkronber

Message:

#2071: merged r13411,r13413,r13414,r13415,r13419,r13440,r13441,r13442,r13445,r13447,r13525,r13526,r13529,r13584,r13901,r13925 from trunk to stable

Location:

stable

Files:

• . (modified) (1 prop)
• HeuristicLab.Tests (modified) (1 prop)
• HeuristicLab.Tests/HeuristicLab.Problems.Instances.DataAnalysis-3.3/TableFileParserTest.cs (modified) (2 diffs)

stable/HeuristicLab.Tests/HeuristicLab.Problems.Instances.DataAnalysis-3.3/TableFileParserTest.cs

@@ -21 +21 @@
 
 using System;
+using System.Collections.Generic;
+using System.Globalization;
 using System.IO;
 using Microsoft.VisualStudio.TestTools.UnitTesting;
@@ -589 +591 @@
     }
 
+
+    [TestMethod]
+    [TestCategory("Problems.Instances")]
+    [TestProperty("Time", "short")]
+    public void ParseWithColumnTypeConversionDE() {
+      // If first entry of a column can be parsed as a double we assume all values are doubles.
+      // However, if any of the following entries cannot be parsed as a double we convert the whole column to a string column.
+      // Special care needs to be taken with missing values, NaN (n.def.) and infinity values.
+      // We only support DE-DE and InvariantCulture number formats
+      string tempFileName = Path.GetTempFileName();
+      var deCultureInfo = CultureInfo.GetCultureInfo("DE-DE");
+      WriteToFile(tempFileName,
+      "str\tdbl\tdbl\tdbl" + Environment.NewLine +
+      "1,3\t1,3\t0\t3" + Environment.NewLine +
+      "1,3\t\t0\t0" + Environment.NewLine +
+      "s\t" + double.NaN.ToString(deCultureInfo) + "\t0\t0" + Environment.NewLine + // double.NaN might have a different string representation on different systems (even when using the same CultureInfo)
+      "s\t" + double.PositiveInfinity.ToString(deCultureInfo) + "\t0\t0" + Environment.NewLine +
+      "s\t" + double.NegativeInfinity.ToString(deCultureInfo) + "\t0\t0" + Environment.NewLine +
+      "s\t0\t0\t0");
+      TableFileParser parser = new TableFileParser();
+      try {
+        parser.Parse(tempFileName,
+          deCultureInfo.NumberFormat,
+          deCultureInfo.DateTimeFormat,
+          '\t',
+          true);
+        Assert.AreEqual(6, parser.Rows);
+        Assert.AreEqual(4, parser.Columns);
+        Assert.IsTrue(parser.Values[0] is List<string>);
+        Assert.IsTrue(parser.Values[1] is List<double>);
+        Assert.IsTrue(parser.Values[2] is List<double>);
+        Assert.IsTrue(parser.Values[3] is List<double>);
+        Assert.IsTrue(double.IsNaN((double)parser.Values[1][1])); // missing value
+        Assert.IsTrue(double.IsNaN((double)parser.Values[1][2]));
+        Assert.IsTrue(double.IsPositiveInfinity((double)parser.Values[1][3])); // NOTE: in DE-DE NumberFormat just "unendlich" is not allowed (compare with InvariantCulture)
+        Assert.IsTrue(double.IsNegativeInfinity((double)parser.Values[1][4]));
+      } finally {
+        File.Delete(tempFileName);
+      }
+    }
+
+    [TestMethod]
+    [TestCategory("Problems.Instances")]
+    [TestProperty("Time", "short")]
+    public void ParseWithColumnTypeConversionInvariant() {
+      // see ParseWithColumnTypeConversionDE above
+      // same routine only using invariant culture
+      string tempFileName = Path.GetTempFileName();
+      WriteToFile(tempFileName,
+      @"str,dbl,dbl,dbl
+1.3,1.3,0,3
+1.3,,0,0
+s,NaN,0,0
+s,Infinity,0,0
+s,-Infinity,0,0
+s,0,0,0");
+      TableFileParser parser = new TableFileParser();
+      try {
+        parser.Parse(tempFileName,
+          CultureInfo.InvariantCulture.NumberFormat,
+          CultureInfo.InvariantCulture.DateTimeFormat,
+          ',',
+          parser.AreColumnNamesInFirstLine(tempFileName));
+        Assert.AreEqual(6, parser.Rows);
+        Assert.AreEqual(4, parser.Columns);
+        Assert.IsTrue(parser.Values[0] is List<string>);
+        Assert.IsTrue(parser.Values[1] is List<double>);
+        Assert.IsTrue(parser.Values[2] is List<double>);
+        Assert.IsTrue(parser.Values[3] is List<double>);
+        Assert.IsTrue(double.IsNaN((double)parser.Values[1][1])); // missing value
+        Assert.IsTrue(double.IsNaN((double)parser.Values[1][2]));
+        Assert.IsTrue(double.IsPositiveInfinity((double)parser.Values[1][3])); // NOTE: in InvariantCulture +Infinity is not allowed (compare with DE-DE)
+        Assert.IsTrue(double.IsNegativeInfinity((double)parser.Values[1][4]));
+      } finally {
+        File.Delete(tempFileName);
+      }
+    }
+
     private void WriteToFile(string fileName, string content) {
       using (StreamWriter writer = new StreamWriter(fileName)) {