trjhtr

I have an IEnumerable of a view model that I need to convert to CSV and send back as a file response. I have written it out quite manually and I'm just seeking code review and suggestions on a more elegant solution, or alternatively if this looks acceptable, confirmation of that:

 public ActionResult ExportBasicReads(string nmi, DateTime fromDate, DateTime toDate)
 
 var reads = _dataService.GetBasicReads(nmi, fromDate, toDate)
 .Where(r => r.CurrentReadQuality != ReadQuality.Estimate)
 .OrderBy(r => r.CurrentReadDate)
 .ThenBy(r => r.DeviceKey)
 .ThenBy(r => r.ValidationCode);

 StringBuilder csv = new StringBuilder();

 csv.AppendLine("Read Date,Device : Suffix,Read,Quality,Tran Code,Service Order,Previous Read Date,Previous Read,Previous Quality," +
 "Previous Tran Code,Dial Diff,Quantity,UOM,Dial Factor,Direction,Validation,From Date,To Date,Kwh,Avg Daily Load,Profile Name," +
 "Profle Area,Reading Timestamp,Update Date,ChangedBy,ChangeType,ChangeComment");

 reads.ForEach(x => csv.AppendLine(
 string.Format("0,1 : 2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22," +
 "23,24,25,26,27",
 x.CurrentReadDate,
 x.DeviceKey,
 x.Suffix,
 x.CurrentReadRawValue,
 x.CurrentReadQualityDesc,
 x.CurrentReadTransactionCode,
 x.CurrentServiceOrder,
 x.PreviousReadDate,
 x.PreviousReadRawValue,
 x.PreviousReadQualityDesc,
 x.PreviousReadTransactionCode,
 x.DialDifference,
 x.Quantity,
 x.UOMDesc,
 x.Multiplier,
 x.DirectionDesc,
 x.ValidationDesc,
 x.EffectiveFromDateTime,
 x.EffectiveToDateTime,
 x.Kwh,
 x.AverageDailyLoad,
 x.ProfileName,
 x.ProfileArea,
 x.EffectiveFromDateTime,
 x.UpdateTimestamp,
 x.ChangedBy,
 x.ChangeType,
 x.ChangeComment
 )
 )); 

 return File(Encoding.ASCII.GetBytes(csv.ToString()), "text/csv", "BasicReadsExport.csv");
 

I modified the formatting part for better readability and maintainability, at the cost of lower performance:

var columns = new Dictionary<string, Func<Reads, object>>

 ["Read Date"] = x => x.CurrentReadDate,
 ["Device"] = x => x.DeviceKey,

 ["Suffix"] = x => x.Suffix,
 ["Read"] = x => x.CurrentReadRawValue,
 // add the rest of headers/property selectors below
;
//var format = "0,1 : 2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27";
var format = "01 : " + string.Join(",", columns.Select((_, i) => $"i").Skip(2));

var csv = new StringBuilder()
 .AppendFormat(format, columns.Keys.ToArray())
 .AppendLine();
foreach (var read in reads)

 var values = columns.Values.Select(selector => selector(read));
 csv
 .AppendFormat(format, values.ToArray())
 .AppendLine();

The advantage of this approach is that you can add/move/remove the columns without having to worry about updating the header and the format, and if they are kept in sync or not.

Note:

• You can also revert to the constant format if it becomes more complex.


• The property selector can be chained if you need special format on individual property, like: x => x.CurrentReadDate.ToString("yyyyMMddThhmm")