尝试读取csv文件时出错

Question

Good Day,

i am having trouble reading csv files on my asp.net project.

it always returns the error index out of range cannot find column 6

before i go on explaining what i did here is the code:

                string savepath;  
                HttpPostedFile postedFile = context.Request.Files["Filedata"];
                savepath = context.Server.MapPath("files");
                string filename = postedFile.FileName;
                todelete = savepath + @"\" + filename;
                string forex = savepath + @"\" + filename;
                postedFile.SaveAs(savepath + @"\" + filename);
                DataTable tblcsv = new DataTable();
                tblcsv.Columns.Add("latitude");
                tblcsv.Columns.Add("longitude");
                tblcsv.Columns.Add("mps");
                tblcsv.Columns.Add("activity_type");
                tblcsv.Columns.Add("date_occured");
                tblcsv.Columns.Add("details");
                string ReadCSV = File.ReadAllText(forex);

                foreach (string csvRow in ReadCSV.Split('\n'))
                {
                    if (!string.IsNullOrEmpty(csvRow))
                    {
                        //Adding each row into datatable  
                        tblcsv.Rows.Add();
                        int count = 0;
                        foreach (string FileRec in csvRow.Split('-'))
                        {
                            tblcsv.Rows[tblcsv.Rows.Count - 1][count] = FileRec;
                            count++;
                        }
                    }


                }

i tried using comma separated columns but the string that comes with it contains comma so i tried the - symbol just to make sure that there are no excess commas on the text file but the same error is popping up.

am i doing something wrong?

thank you in advance

Answer 1

Your excel file might have more columns than 6 for one or more rows. For this reason the splitting in inner foreach finds more columns but the tblcsv does not have more columns than 6 to assign the extra column value.

Try something like this:

foreach (string FileRec in csvRow.Split('-'))
{
    if(count > 5)
        return;
    tblcsv.Rows[tblcsv.Rows.Count - 1][count] = FileRec;
    count++;
}

However it would be better if you check for additional columns before processing and handle the issue.

Answer 2

StringBuilder errors = new StringBuilder();   //// this will hold the record for those array which have length greater than the 6

foreach (string csvRow in ReadCSV.Split('\n'))
                {
                    if (!string.IsNullOrEmpty(csvRow))
                    {
                        //Adding each row into datatable  
                       DataRow dr = tblcsv.NewRow(); and then  
                        int count = 0;
                        foreach (string FileRec in csvRow.Split('-'))
                        {
                     try
                      {
                           dr[count] = FileRec;
                            tblcsv.Rows.Add(dr);
                      }
                     catch (IndexOutOfRangeException i)
                     {
                     error.AppendLine(csvRow;)
                     break;
                     }
                            count++;
                    }
                    }
                }

Now in this case we will have the knowledge of the csv row which is causing the errors, and rest will be processed successfully. Validate the row in errors whether its desired input, if not then correct value in csv file.

Answer 3

You can't treat the file as a CSV if the delimiter appears inside a field. In this case you can use a regular expression to extract the first five fields up to the dash, then read the rest of the line as the sixth field. With a regex you can match the entire string and even avoid splitting lines.

Regular expressions are also a lot faster than splits and consume less memory because they don't create temporary strings. That's why they are used extensively to parse log files. The ability to capture fields by name doesn't hurt either

The following sample parses the entire file and captures each field in a named group. The last field captures everything to the end of the line:

var pattern="^(?<latitude>.*?)-(?<longitude>.*?)-(?<mps>.*?)-(?<activity_type>.*?)-" +
            "(?<date_occured>.*?)-(?<detail>.*)$";
var regex=new Regex(pattern,RegexOptions.Multiline);
var matches=regex.Matches(forex);

foreach (Match match in matches)
{
    DataRow dr = tblcsv.NewRow();
    row["latitude"]=match.Groups["latitude"].Value);
    row["longitude"]=match.Groups["longitude"].Value);
    ...
    tblcsv.Rows.Add(dr);
}

The (?<latitude>.*?)- pattern captures everything up to the first dash into a group named latitude . The .*? pattern means the matching isn't greedy ie it won't try to capture everything to the end of the line but will stop when the first - is encountered.

The column names match the field names, which means you can add all fields with a loop:

foreach (Match match in matches)
{
    var row = tblCsv.NewRow();
    foreach (Group group in match.Groups)
    {
        foreach (DataColumn col in tblCsv.Columns)
        {
            row[col.ColumnName]=match.Groups[col.ColumnName].Value;                        
        }                    
    }
    tblCsv.Rows.Add(row);
}
tblCsv.Rows.Add(row);