Click here to Skip to main content
Click here to Skip to main content

Pivoting DataTable Simplified

, 20 Aug 2014 CPOL
Rate this:
Please Sign up or sign in to vote.
A class to pivot a DataTable with various aggregate functions.

Introduction

Displaying data in tabular form is an essential part of any application nowadays. But sometimes you need to display a huge amount of data in terms of number of rows. It becomes very difficult to analyse if the number of rows is huge. In such cases, you may wish to summarize your data in the other formats like charts, graphs, groups, pivots, etc. This article presents a simplified way to pivot your data with an appropriate aggregate function so that you can enhance your reports easily. Based on the feedback given by various readers, more features have been provided to pivot class. The pivot class is now capable to pivot data on both the axis at a time. Moreover, it also has the facility to do sub-total column wise.  

Below is a screenshot of pivoted data in a GridView

PivotDataTable/2ndPivot.JPG

How it Works

To simplify the scenario, I have divided the result table into three areas: RowField, DataField, and ColumnFields. If you wish to do pivot on both the axis, you may use another overload of the same method where you just need to pass RowFields parameter as an array. Apart from the area, the Pivot class provides you the option to bind your data based on some aggregate functions. The various aggregate options available are:

  • Count: Returns the count of matching data
  • Sum: Returns the sum of matching data (to get the sum, the type of the DataField must be convertible to decimal type)
  • First: Returns the first occurrence of matching data
  • Last: Returns the last occurrence of matching data
  • Average: Returns the average of matching data (to get the average, the type of the DataField must be convertible to decimal type)
  • Max: Returns the maximum value from the matching data
  • Min: Returns the minimum value from the matching data
  • Exists: Returns "true" if there is any matching data, else "false"

The code mainly contains a class named "Pivot" that takes the DataTable in the constructor. ColumnFields takes as a string array parameter which allows you to pivot data on more than one column. It contains a function called PivotData() which actually pivots your data.

public DataTable PivotData(string RowField, string DataField, 
       AggregateFunction Aggregate, params string[] ColumnFields)
{
    DataTable dt = new DataTable();
    string Separator = ".";
    var RowList = (from x in _SourceTable.AsEnumerable() 
        select new { Name = x.Field<object>(RowField) }).Distinct();
    var ColList = (from x in _SourceTable.AsEnumerable() 
                   select new { Name = ColumnFields.Select(n => x.Field<object>(n))
                       .Aggregate((a, b) => a += Separator + b.ToString()) })
                       .Distinct()
                       .OrderBy(m => m.Name);

    dt.Columns.Add(RowField);
    foreach (var col in ColList)
    {
        dt.Columns.Add(col.Name.ToString());
    }

    foreach (var RowName in RowList)
    {
        DataRow row = dt.NewRow();
        row[RowField] = RowName.Name.ToString();
        foreach (var col in ColList)
        {
            string strFilter = RowField + " = '" + RowName.Name + "'";
            string[] strColValues = 
              col.Name.ToString().Split(Separator.ToCharArray(), 
                                        StringSplitOptions.None);
            for (int i = 0; i < ColumnFields.Length; i++)
                strFilter += " and " + ColumnFields[i] + 
                             " = '" + strColValues[i] + "'";
            row[col.Name.ToString()] = GetData(strFilter, DataField, Aggregate);
        }
        dt.Rows.Add(row);
    }
    return dt;
}

PivotData method also has 2 more overloads. If you wish to show column wise sub-total, you may use the overload by passing a bool variable showSubTotal. If you wish to Pivot your data on both side, i.e., row-wise as well as column-wise, you may wish to use another overload where you can pass rowFields and columnFields as an array. 

First of all, the function determines the number of rows by getting the distinct values in RowList, and the number of columns by getting the distinct values in ColList. Then, the columns are created. It then iterates through each row and gets the matching values to the corresponding cell based on the aggregate function provided. To retrieve the matching value, the GetData() function is called.

private object GetData(string Filter, string DataField, AggregateFunction Aggregate)
{
    try
    {
        DataRow[] FilteredRows = _SourceTable.Select(Filter);
        object[] objList = 
         FilteredRows.Select(x => x.Field<object>(DataField)).ToArray();

        switch (Aggregate)
        {
            case AggregateFunction.Average:
                return GetAverage(objList);
            case AggregateFunction.Count:
                return objList.Count();
            case AggregateFunction.Exists:
                return (objList.Count() == 0) ? "False" : "True";
            case AggregateFunction.First:
                return GetFirst(objList);
            case AggregateFunction.Last:
                return GetLast(objList);
            case AggregateFunction.Max:
                return GetMax(objList);
            case AggregateFunction.Min:
                return GetMin(objList);
            case AggregateFunction.Sum:
                return GetSum(objList);
            default:
                return null;
        }
    }
    catch (Exception ex)
    {
        return "#Error";
    }
    return null;
}

This function first filters out the matching RowField and ColumnFields data in the DataRow[] array and then applies the aggregate function on it.

Using the Code

Using the code is simple. Create an instance of the Pivot class and then call the PivotData method with the required parameters. The PivotData() method returns the DataTable which can directly be used as the DataSource of the GridView

DataTable dt = ExcelLayer.GetDataTable("_Data\\DataForPivot.xls", "Sheet1$");
Pivot pvt = new Pivot(dt);

grdPivot.DataSource = pvt.PivotData("Designation", "CTC", 
   AggregateFunction.Max, "Company", "Department", "Year");
grdPivot.DataBind();

The database used as a sample is an Excel sheet and is present in the "_Data" folder of the root folder of sample application.

Merge GridView Header Cells

The MergeHeader function is created to merge the header cells to provide a simplified look.

private void MergeHeader(GridView gv, GridViewRow row, int PivotLevel)
{
    for (int iCount = 1; iCount <= PivotLevel; iCount++)
    {
        GridViewRow oGridViewRow = new GridViewRow(0, 0, 
          DataControlRowType.Header, DataControlRowState.Insert);
        var Header = (row.Cells.Cast<tablecell>()
            .Select(x => GetHeaderText(x.Text, iCount, PivotLevel)))
            .GroupBy(x => x);

        foreach (var v in Header)
        {
            TableHeaderCell cell = new TableHeaderCell();
            cell.Text = v.Key.Substring(v.Key.LastIndexOf(_Separator) + 1);
            cell.ColumnSpan = v.Count();
            oGridViewRow.Cells.Add(cell);
        }
        gv.Controls[0].Controls.AddAt(row.RowIndex, oGridViewRow);
    }
    row.Visible = false;
}

The function creates a new row for each pivot level and merges accordingly. PivotLevel here is the number of columns on which the pivot is done.

Header gets all the column values in an array, groups the repeated values returned by the GetHeaderText() function, sets the ColumnSpan property of the newly created cell according to the number of repeated HeaderText, and then adds the cell to the GridViewRow. Finally, add the GridViewRow to the GridView.

The GetHeaderText() function returns the header text based on the PivotLevel.

For example, suppose a pivot is done on three ColumnFields, namely, Company, Department, and Year. The result header of the GridView will initially have a header like Company.Department.Year for a PivotLevel 1. GetHeaderText() will return Company. For a PivotLevel 2, GetHeaderText() will return Company.Department. For a PivotLevel 3, GetHeaderText() will return Company.Department.Year, and so on... 

Merge GridView Row Header Cells

This may needs to be done when you are pivoting your data row-wise also. Here we are simply merging the cells with same text.

private void MergeRows(GridView gv, int rowPivotLevel)
    {
        for (int rowIndex = gv.Rows.Count - 2; rowIndex >= 0; rowIndex--)
        {
            GridViewRow row = gv.Rows[rowIndex];
            GridViewRow prevRow = gv.Rows[rowIndex + 1];
            for (int colIndex = 0; colIndex < rowPivotLevel; colIndex++)
            {
                if (row.Cells[colIndex].Text == prevRow.Cells[colIndex].Text)
                {
                    row.Cells[colIndex].RowSpan = (prevRow.Cells[colIndex].RowSpan < 2) ? 2 : prevRow.Cells[colIndex].RowSpan + 1;
                    prevRow.Cells[colIndex].Visible = false;
                }
            }
        }
    }

The code to merge header rows is fairly simple. It simply loops through all the row header cells from bottom to top, compare the text with previous corresponding row cell, increases the row span by 1 if same and hide the previous corresponding row.

Screen shot for both side pivot:

Both Side Pivot

Below is the screenshot of the GridView containing the third level pivoted data:

PivotDataTable/3rdPivot.JPG

Points of Interest

Along with pivoting the DataTable, the code will also help you to merge the header cells in the desired format in GridView. Moreover, you may have a deeper look into PivotData method to know how you can search or filter data in DataTable suing linq. Apart from this, MergeRows method acts as a sample to merge rows in a GridView. For beginners, the ExcelLayer.GetDataTable() method will be a sample to get the data from the Excel Sheet.

Based on the request from many readers, i have not provided the sample to query the data from database too. You may find the Sql Script attached to create Sql server database table and code to read data to DataTable from Sql Server.

You may also wish to consider the following link to pivot a DataTable: http://www.codeproject.com/KB/recipes/CsharpPivotTable.aspx.

Future Consideration

Currently, the code can pivot data only for a DataTable. The code will be enhanced to pivot any object derived from an IListSource or ICollection.

History

  • First version release: December 09, 2009.
  • VB.NET source and demo added: March 19, 2010.

License

This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Share

About the Author

Anurag Gandhi
Technical Lead
India India
Anurag Gandhi currently works in web application design/development and has done so for many years now as he is passionate about programming.
He is extensively involved in Asp.Net and Asp.Net MVC web application architecture, AngularJs, design and development. His languages of choice are C#, Asp.Net, MVC, Asp, C, C++. But he is familiar with many other coding languages. He works with MS Sql Server as the database architecture of choice.
He is active in programming communities and loves to share his knowledge with other developers whenever he gets the opportunity.
He is also a passionate chess player.
 
He can be contacted at: soft.gandhi@gmail.com
Follow on   Twitter   Google+

Comments and Discussions

 
Questionimporting into VS 2013 incompatability issue Pinmemberrockinvet22-Jan-14 8:35 
AnswerRe: importing into VS 2013 incompatability issue PinmemberAnurag Gandhi25-Jan-14 13:33 
QuestionI get the error shown below PinmemberMember 103712024-Nov-13 15:38 
AnswerRe: I get the error shown below PinmemberAnurag Gandhi4-Nov-13 20:49 
QuestionLarge Datasets ? [modified] PinmemberMember 31019414-Nov-13 11:49 
AnswerRe: Large Datasets ? PinmemberAnurag Gandhi4-Nov-13 20:47 
QuestionWhat about multiple data-fields? PinprofessionalAnoop Ananthan7-Oct-13 21:47 
AnswerRe: What about multiple data-fields? PinmemberAnurag Gandhi7-Oct-13 22:55 
QuestionThank you! Pinmembernordman2510-Sep-13 6:26 
AnswerRe: Thank you! PinmemberAnurag Gandhi11-Sep-13 6:42 
GeneralMy vote of 5 Pinmembersusanna.floora5-Sep-13 3:58 
GeneralRe: My vote of 5 PinmemberAnurag Gandhi11-Sep-13 6:41 
QuestionThank you and quick question. Pinmembersykiemikey3-Aug-13 10:00 
AnswerRe: Thank you and quick question. PinmemberAnurag Gandhi4-Aug-13 8:23 
QuestionImplement IDisposable or don't cache DataTable Pinmemberjohnmcalvert30-Jul-13 8:21 
AnswerRe: Implement IDisposable or don't cache DataTable PinmemberAnurag Gandhi31-Jul-13 8:43 
QuestionHeader name PinmemberMember 990188818-Jun-13 4:43 
AnswerRe: Header name PinmemberAnurag Gandhi18-Jun-13 6:18 
Questionquestion regarding Both side pivot PinmemberAdorable15-Jun-13 8:10 
AnswerRe: question regarding Both side pivot PinmemberAdorable15-Jun-13 9:03 
AnswerRe: question regarding Both side pivot PinmemberAnurag Gandhi17-Jun-13 22:04 
GeneralRe: question regarding Both side pivot Pinmemberroberto galbiati25-Aug-14 0:30 
Questionbroken link in Points of Interest Pinmemberjohnmcalvert14-May-13 7:31 
AnswerRe: broken link in Points of Interest PinmemberAnurag Gandhi17-May-13 21:51 
QuestionOrder column PinmemberMember 990188811-Mar-13 11:59 
AnswerRe: Order column Pinmemberjohnmcalvert16-May-13 12:06 
AnswerRe: Order column PinmemberAnurag Gandhi17-May-13 22:14 
QuestionThank you Anurag Gandhi [modified] Pinmemberarulb2w8-Mar-13 20:39 
AnswerRe: Thank you Anurag Gandhi PinmemberAnurag Gandhi9-Mar-13 2:19 
QuestionCouple of problems... Pinmembercjard4-Feb-13 1:43 
AnswerRe: Couple of problems... PinmemberAnurag Gandhi8-Feb-13 17:57 
GeneralMy vote of 5 PinmemberMember 760623810-Jan-13 7:31 
GeneralRe: My vote of 5 PinmemberAnurag Gandhi8-Feb-13 17:58 
Questionthanks Pinmemberaj18830-Dec-12 4:09 
AnswerRe: thanks Pinmemberaj18830-Dec-12 11:59 
GeneralGood Work PinmemberMember 790590310-Dec-12 22:32 
GeneralRe: Good Work PinmemberAnurag Gandhi19-Dec-12 20:57 
QuestionWhat about Entity framework Pinmemberjoker220627-Nov-12 4:08 
AnswerRe: What about Entity framework PinmemberAnurag Gandhi17-May-13 22:16 
QuestionReport on two values? Pinmembergeezer9917-Nov-12 17:11 
QuestionElegant, had to modify a bit to handle translation of DataTypes to new table Pinmemberdternes5-Nov-12 7:39 
AnswerRe: Elegant, had to modify a bit to handle translation of DataTypes to new table PinmemberAnurag Gandhi5-Nov-12 23:11 
GeneralRe: Elegant, had to modify a bit to handle translation of DataTypes to new table [modified] Pinmemberjohnmcalvert14-May-13 7:26 
GeneralRe: Elegant, had to modify a bit to handle translation of DataTypes to new table PinmemberAnurag Gandhi24-May-13 21:20 
QuestionHelp with Importing Pivot PinmemberHyprHare31-Oct-12 5:50 
AnswerRe: Help with Importing Pivot PinmemberAnurag Gandhi1-Nov-12 10:34 
QuestionPlease add complete VB.net code Pinmemberezar2312-Sep-12 13:15 
AnswerRe: Please add complete VB.net code PinmemberAnurag Gandhi13-Sep-12 22:40 
QuestionError using VS 2008 PinmemberLeoLisogorsky3-Aug-12 9:17 
I open the project and can't Build. Get error "Unrecognized attribute 'targetFranmework'.
 
If I run last build, what I get instead is:
 
Server Error in '/PivotDataTable' Application.
--------------------------------------------------------------------------------
 
Configuration Error
Description: An error occurred during the processing of a configuration file required to service this request. Please review the specific error details below and modify your configuration file appropriately.
 
Parser Error Message: Unrecognized attribute 'targetFramework'. Note that attribute names are case-sensitive.
 
Source Error:
 

Line 18:             during development.
Line 19:         -->
 
Thanks for help,
Leo.
GeneralRe: Error using VS 2008 PinmemberAnurag Gandhi5-Aug-12 8:06 

General General    News News    Suggestion Suggestion    Question Question    Bug Bug    Answer Answer    Joke Joke    Rant Rant    Admin Admin   

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.

| Advertise | Privacy | Terms of Use | Mobile
Web02 | 2.8.1411022.1 | Last Updated 20 Aug 2014
Article Copyright 2009 by Anurag Gandhi
Everything else Copyright © CodeProject, 1999-2014
Layout: fixed | fluid