CSV Reader delimiter problem

Question

Hi!

I have a script that reads a csv file and generates data in order to instantiate a prefab. I'm currently using ';' as delimiter to separate, but if there is a ";" in the string it jumps into the next column. Any idea on how I can have a ";" in a string and it doesn't go to the next column? This is the code:

 [SerializeField] DataScript dataObject;

 private void Start ()
 {
     CreateDataObjects();
 }
 
 private void CreateDataObjects ()
 {
     StreamReader strReader = new StreamReader(Path.GetFullPath("dataObjectsSheet.csv"));
     strReader.ReadLine();
     bool endOfFile = false;
     while (!endOfFile && (strReader.Peek() != -1))
     {
         string data_String = strReader.ReadLine();
         if (data_String == null)
         {
             endOfFile = true;
             break;
         }
         var dataValues = data_String.Split(';');
         dataObject.name = dataValues[0].ToString();
         dataObject.month = float.Parse(dataValues[1].ToString());
         dataObject.ranking = int.Parse(dataValues[2].ToString());
         dataObject.days = int.Parse(dataValues[3].ToString());
         dataObject.nick = dataValues[4].ToString();
         dataObject.bio = dataValues[5].ToString();
         dataObject.url = dataValues[6].ToString();
         Instantiate(dataObject, transform);
     }
 }

Answer 1

Well, the classical CSV format would simply put a value in double quotes if i contains a delimiter, double quote or new line character. Of course you can not simply split a line at the seperation character and you have to actually parse through the text, character by character. That's what string.split does anyways internally. Even CSV is a simple format it still has some rules you have to follow. Simply splitting the string at a delimiter does work for trivial cases but it does not work for the actual specified CSV format.

I've quickly created this CSVParser which you can use like this:

 using (StreamReader strReader = new StreamReader(Path.GetFullPath("dataObjectsSheet.csv")))
 {
     CSVParser parser = new CSVParser(strReader, ';');
     var line = new List<string>();
     parser.NextLine(line); // skip header line
     while (parser.NextLine(line))
     {
         // use line[0] to access the first "value", line[1] for the second, ...
     }
 }

Keep in mind that a line in the CSV file could potentially contain different number of values / columns. Blindly reading say line[5] would break if a line doesn't contain at least 6 values. So further parsing the values might need some additional checks.

Also note that float.Parse without a specific culture will use the system culture. So if this code should run on a user's machine, make sure the user uses the right culture or, if needed, use a specific culture. The most common culture is the invariant culture since it won't change in the future. Country specific cultures can change in the future which could break your program. For example my local culture is german so our decimal delimiter is the comma while in the english speaking world the dot is usually used. Note that there are quite a few countries which use the comma

CSV Reader delimiter problem

1 Reply

Your answer

Follow this Question

Related Questions