Related
"0.0.0.0,""0.255.255.255"",""ZZ"""
"1.0.0.0,""1.0.0.255"",""AU"""
"1.0.1.0,""1.0.3.255"",""CN"""
"1.0.4.0,""1.0.7.255"",""AU"""
"1.0.8.0,""1.0.15.255"",""CN"""
"1.0.16.0,""1.0.31.255"",""JP"""
"1.0.32.0,""1.0.63.255"",""CN"""
"1.0.64.0,""1.0.127.255"",""JP"""
"1.0.128.0,""1.0.255.255"",""TH"""
"1.1.0.0,""1.1.0.255"",""CN"""
"1.1.1.0,""1.1.1.255"",""AU"""
"1.1.2.0,""1.1.63.255"",""CN"""
"1.1.64.0,""1.1.127.255"",""JP"""
"1.1.128.0,""1.1.255.255"",""TH"""
İN EXCEL
0.0.0.0,"0.255.255.255","ZZ"
1.0.0.0,"1.0.0.255","AU"
1.0.1.0,"1.0.3.255","CN"
1.0.4.0,"1.0.7.255","AU"
1.0.8.0,"1.0.15.255","CN"
1.0.16.0,"1.0.31.255","JP"
1.0.32.0,"1.0.63.255","CN"
1.0.64.0,"1.0.127.255","JP"
1.0.128.0,"1.0.255.255","TH"
1.1.0.0,"1.1.0.255","CN"
1.1.1.0,"1.1.1.255","AU"
1.1.2.0,"1.1.63.255","CN"
1.1.64.0,"1.1.127.255","JP"
1.1.128.0,"1.1.255.255","TH"
1.2.0.0,"1.2.2.255","CN"
1.2.3.0,"1.2.3.255","AU"
1.2.4.0,"1.2.127.255","CN"
1.2.128.0,"1.2.255.255","TH"
1.3.0.0,"1.3.255.255","CN"
1.4.0.0,"1.4.0.255","AU"
1.4.1.0,"1.4.127.255","CN"
1.4.128.0,"1.4.255.255","TH"
How can split this CSV file.
For example 0.0.0.0 0.255.255.255 ZZ for first row and how can add datagridview with 3columns
You can do it via the following way..
using System.IO;
static void Main(string[] args)
{
using(var reader = new StreamReader(#"C:\test.csv"))
{
List<string> listA = new List<string>();
List<string> listB = new List<string>();
while (!reader.EndOfStream)
{
var line = reader.ReadLine();
var values = line.Split(','); // or whatever yur get by reading that file
listA.Add(values[0]);
listB.Add(values[1]);
}
}
}
A CSV file is either a Tab delimited or a Comma delimited file. That said; you have to read the file line by line and then separate the values available in a line based on the delimiter character. The first line usually appears in a CSV file is usually the headers which you can use in order to produce a KeyValue pair to make your collection more efficient. For example:
Dictionary<int, Dictionary<String, String>> values = new Dictionary<int, Dictionary<String,String>>();
using(FileStream fileStream = new FileStream(#"D:\MyCSV.csv", FileMode.Open, FileAccess.Read, FileShare.Read)) {
using(StreamReader streamReader = new StreamReader(fileStream)){
//You can skip this line if there is no header
// Then instead of Dictionary<String,String> you use List<String>
var headers = streamReader.ReadLine().Split(',');
String line = null;
int lineNumber = 1;
while(!streamReader.EndOfStream){
line = streamReader.ReadLine().split(',');
if(line.Length == headers.Length){
var temp = new Dictionary<String, String>();
for(int i = 0; i < headers.Length; i++){
// You can remove '"' character by line[i].Replace("\"", "") or through using the Substring method
temp.Add(headers[i], line[i]);
}
values.Add(lineNumber, temp);
}
lineNumber++;
}
}
In case the data structure of your CSV is constant and it will not change in the future, you can develop a strongly typed data model and get rid of the Dictionary type. This approach will be more elegant and more efficient.
First of all, your CSV lines are surrounded by quotes. Is it copy/paste mistake? If not, you will need to sanitize the file to a valid CSV file.
You can try Cinchoo ETL - an open source library to load the CSV file to datatable, then you can assign it to your DataGridView source.
I'll show you both approach, how to handle
Valid CSV: (test.csv)
0.0.0.0,"0.255.255.255","ZZ"
1.0.0.0,"1.0.0.255","AU"
1.0.1.0,"1.0.3.255","CN"
1.0.4.0,"1.0.7.255","AU"
1.0.8.0,"1.0.15.255","CN"
1.0.16.0,"1.0.31.255","JP"
1.0.32.0,"1.0.63.255","CN"
1.0.64.0,"1.0.127.255","JP"
1.0.128.0,"1.0.255.255","TH"
1.1.0.0,"1.1.0.255","CN"
1.1.1.0,"1.1.1.255","AU"
1.1.2.0,"1.1.63.255","CN"
1.1.64.0,"1.1.127.255","JP"
1.1.128.0,"1.1.255.255","TH"
Read CSV:
using (var p = new ChoCSVReader("test.csv"))
{
var dt = p.AsDataTable();
//Assign dt to DataGridView
}
Next approach
Invalid CSV: (test.csv)
"0.0.0.0,""0.255.255.255"",""ZZ"""
"1.0.0.0,""1.0.0.255"",""AU"""
"1.0.1.0,""1.0.3.255"",""CN"""
"1.0.4.0,""1.0.7.255"",""AU"""
"1.0.8.0,""1.0.15.255"",""CN"""
"1.0.16.0,""1.0.31.255"",""JP"""
"1.0.32.0,""1.0.63.255"",""CN"""
"1.0.64.0,""1.0.127.255"",""JP"""
"1.0.128.0,""1.0.255.255"",""TH"""
"1.1.0.0,""1.1.0.255"",""CN"""
"1.1.1.0,""1.1.1.255"",""AU"""
"1.1.2.0,""1.1.63.255"",""CN"""
"1.1.64.0,""1.1.127.255"",""JP"""
"1.1.128.0,""1.1.255.255"",""TH"""
Read CSV:
using (var p = new ChoCSVReader("Sample6.csv"))
{
p.SanitizeLine += (o, e) =>
{
string line = e.Line as string;
if (line != null)
{
line = line.Substring(1, line.Length - 2);
line = line.Replace(#"""""", #"""");
}
e.Line - line;
};
var dt = p.AsDataTable();
//Assign dt to DataGridView
}
Hope it helps.
i am currently working on a small Project and i got stuck with a Problem i currently can not manage to solve...
I have multiple ".CSV" Files i want to read, they all have the same Data just with different Values.
Header1;Value1;Info1
Header2;Value2;Info2
Header3;Value3;Info3
While reading the first File i Need to Create the Headers. The Problem is they are not splited in Columns but in rows (as you can see above Header1-Header3).
Then it Needs to read the Value 1 - Value 3 (they are listed in the 2nd Column) and on top of that i Need to create another Header -> Header4 with the data of "Info2" which is always placed in Column 3 and Row 2 (the other values of Column 3 i can ignore).
So the Outcome after the first File should look like this:
Header1;Header2;Header3;Header4;
Value1;Value2;Value3;Info2;
And after multiple files it sohuld be like this:
Header1;Header2;Header3;Header4;
Value1;Value2;Value3;Value4;
Value1b;Value2b;Value3b;Value4b;
Value1c;Value2c;Value3c;Value4c;
I tried it with OleDB but i get the Error "missing ISAM" which i cant mange to fix. The Code i Used is the following:
public DataTable ReadCsv(string fileName)
{
DataTable dt = new DataTable("Data");
/* using (OleDbConnection cn = new OleDbConnection("Provider=Microsoft.Jet.OLEDB.4.0;Data Source=\"" +
Path.GetDirectoryName(fileName) + "\";Extendet Properties ='text;HDR=yes;FMT=Delimited(,)';"))
*/
using (OleDbConnection cn = new OleDbConnection("Provider=Microsoft.Jet.OLEDB.4.0;Data Source=" +
Path.GetDirectoryName(fileName) + ";Extendet Properties ='text;HDR=yes;FMT=Delimited(,)';"))
{
using(OleDbCommand cmd = new OleDbCommand(string.Format("select *from [{0}]", new FileInfo(fileName).Name,cn)))
{
cn.Open();
using(OleDbDataAdapter adapter = new OleDbDataAdapter(cmd))
{
adapter.Fill(dt);
}
}
}
return dt;
}
Another attempt i did was using StreamReader. But the Headers are in the wrong place and i dont know how to Change this + do this for every file. the Code i tried is the following:
public static DataTable ReadCsvFilee(string path)
{
DataTable oDataTable = new DataTable();
var fileNames = Directory.GetFiles(path);
foreach (var fileName in fileNames)
{
//initialising a StreamReader type variable and will pass the file location
StreamReader oStreamReader = new StreamReader(fileName);
// CONTROLS WHETHER WE SKIP A ROW OR NOT
int RowCount = 0;
// CONTROLS WHETHER WE CREATE COLUMNS OR NOT
bool hasColumns = false;
string[] ColumnNames = null;
string[] oStreamDataValues = null;
//using while loop read the stream data till end
while (!oStreamReader.EndOfStream)
{
String oStreamRowData = oStreamReader.ReadLine().Trim();
if (oStreamRowData.Length > 0)
{
oStreamDataValues = oStreamRowData.Split(';');
//Bcoz the first row contains column names, we will poluate
//the column name by
//reading the first row and RowCount-0 will be true only once
// CHANGE TO CHECK FOR COLUMNS CREATED
if (!hasColumns)
{
ColumnNames = oStreamRowData.Split(';');
//using foreach looping through all the column names
foreach (string csvcolumn in ColumnNames)
{
DataColumn oDataColumn = new DataColumn(csvcolumn.ToUpper(), typeof(string));
//setting the default value of empty.string to newly created column
oDataColumn.DefaultValue = string.Empty;
//adding the newly created column to the table
oDataTable.Columns.Add(oDataColumn);
}
// SET COLUMNS CREATED
hasColumns = true;
// SET RowCount TO 0 SO WE KNOW TO SKIP COLUMNS LINE
RowCount = 0;
}
else
{
// IF RowCount IS 0 THEN SKIP COLUMN LINE
if (RowCount++ == 0) continue;
//creates a new DataRow with the same schema as of the oDataTable
DataRow oDataRow = oDataTable.NewRow();
//using foreach looping through all the column names
for (int i = 0; i < ColumnNames.Length; i++)
{
oDataRow[ColumnNames[i]] = oStreamDataValues[i] == null ? string.Empty : oStreamDataValues[i].ToString();
}
//adding the newly created row with data to the oDataTable
oDataTable.Rows.Add(oDataRow);
}
}
}
//close the oStreamReader object
oStreamReader.Close();
//release all the resources used by the oStreamReader object
oStreamReader.Dispose();
}
return oDataTable;
}
I am thankful for everyone who is willing to help. And Thanks for reading this far!
Sincerely yours
If I understood you right, there is a strict parsing there like this:
string OpenAndParse(string filename, bool firstFile=false)
{
var lines = File.ReadAllLines(filename);
var parsed = lines.Select(l => l.Split(';')).ToArray();
var header = $"{parsed[0][0]};{parsed[1][0]};{parsed[2][0]};{parsed[1][0]}\n";
var data = $"{parsed[0][1]};{parsed[1][1]};{parsed[2][1]};{parsed[1][2]}\n";
return firstFile
? $"{header}{data}"
: $"{data}";
}
Where it would return - if first file:
Header1;Header2;Header3;Header2
Value1;Value2;Value3;Value4
if not first file:
Value1;Value2;Value3;Value4
If I am correct, rest is about running this against a list file of files and joining the results in an output file.
EDIT: Against a directory:
void ProcessFiles(string folderName, string outputFileName)
{
bool firstFile = true;
foreach (var f in Directory.GetFiles(folderName))
{
File.AppendAllText(outputFileName, OpenAndParse(f, firstFile));
firstFile = false;
}
}
Note: I missed you want a DataTable and not an output file. Then you could simply create a list and put the results into that list making the list the datasource for your datatable (then why would you use semicolons in there? Probably all you need is to simply attach the array values to a list).
(Adding as another answer just to make it uncluttered)
void ProcessMyFiles(string folderName)
{
List<MyData> d = new List<MyData>();
var files = Directory.GetFiles(folderName);
foreach (var file in files)
{
OpenAndParse(file, d);
}
string[] headers = GetHeaders(files[0]);
DataGridView dgv = new DataGridView {Dock=DockStyle.Fill};
dgv.DataSource = d;
dgv.ColumnAdded += (sender, e) => {e.Column.HeaderText = headers[e.Column.Index];};
Form f = new Form();
f.Controls.Add(dgv);
f.Show();
}
string[] GetHeaders(string filename)
{
var lines = File.ReadAllLines(filename);
var parsed = lines.Select(l => l.Split(';')).ToArray();
return new string[] { parsed[0][0], parsed[1][0], parsed[2][0], parsed[1][0] };
}
void OpenAndParse(string filename, List<MyData> d)
{
var lines = File.ReadAllLines(filename);
var parsed = lines.Select(l => l.Split(';')).ToArray();
var data = new MyData
{
Col1 = parsed[0][1],
Col2 = parsed[1][1],
Col3 = parsed[2][1],
Col4 = parsed[1][2]
};
d.Add(data);
}
public class MyData
{
public string Col1 { get; set; }
public string Col2 { get; set; }
public string Col3 { get; set; }
public string Col4 { get; set; }
}
I don't know if this is the best way to do this. But what i would have done in your case, is to rewrite the CSV's the conventionnal way while reading all the files, then create a stream containing the new CSV created.
It would look like something like this :
var csv = new StringBuilder();
csv.AppendLine("Header1;Header2;Header3;Header4");
foreach (var item in file)
{
var newLine = string.Format("{0},{1},{2},{3}", item.value1, item.value2, item.value3, item.value4);
csv.AppendLine(newLine);
}
//Create Stream
MemoryStream stream = new MemoryStream();
StreamReader reader = new StreamReader(stream);
//Fill your data table here with your values
Hope this will help.
i have to import 2 CSV's.
CSV 1 [49]: Including about 50 tab seperated colums.
CSV 2:[2] Inlcudes 3 Columns which should be replaced on the [3] [6] and [11] place of my first csv.
So heres what i do:
1) Importing the csv and split into a array.
string employeedatabase = "MYPATH";
List<String> status = new List<String>();
StreamReader file2 = new System.IO.StreamReader(filename);
string line = file2.ReadLine();
while ((line = file2.ReadLine()) != null)
{
string[] ud = line.Split('\t');
status.Add(ud[0]);
}
String[] ud_status = status.ToArray();
PROBLEM 1: i have about 50 colums to handle, ud_status is just the first, so do i need 50 Lists and 50 String arrays?
2) Importing the second csv and split into a array.
List<String> vorname = new List<String>();
List<String> nachname = new List<String>();
List<String> username = new List<String>();
StreamReader file = new System.IO.StreamReader(employeedatabase);
string line3 = file.ReadLine();
while ((line3 = file.ReadLine()) != null)
{
string[] data = line3.Split(';');
vorname.Add(data[0]);
nachname.Add(data[1]);
username.Add(data[2]);
}
String[] db_vorname = vorname.ToArray();
String[] db_nachname = nachname.ToArray();
String[] db_username = username.ToArray();
PROBLEM 2: After loading these two csv's i dont know how to combine them, and change to columns as mentioned above ..
somethine like this?
mynewArray = ud_status + "/t" + ud_xy[..n] + "/t" + changed_colum + ud_xy[..n];
save "mynewarray" into tablulator seperated csv with encoding "utf-8".
To read the file into a meaningful format, you should set up a class that defines the format of your CSV:
public class CsvRow
{
public string vorname { get; set; }
public string nachname { get; set; }
public string username { get; set; }
public CsvRow (string[] data)
{
vorname = data[0];
nachname = data[1];
username = data[2];
}
}
Then populate a list of this:
List<CsvRow> rows = new List<CsvRow>();
StreamReader file = new System.IO.StreamReader(employeedatabase);
string line3 = file.ReadLine();
while ((line3 = file.ReadLine()) != null)
{
rows.Add(new CsvRow(line3.Split(';'));
}
Similarly format your other CSV and include unused properties for the new fields. Once you have loaded both, you can populate the new properties from this list in a loop, matching the records by whatever common field the CSVs hopefully share. Then finally output the resulting data to a new CSV file.
Your solution is not to use string arrays to do this. That will just drive you crazy. It's better to use the System.Data.DataTable object.
I didn't get a chance to test the LINQ lambda expression at the end of this (or really any of it, I wrote this on a break), but it should get you on the right track.
using (var ds = new System.Data.DataSet("My Data"))
{
ds.Tables.Add("File0");
ds.Tables.Add("File1");
string[] line;
using (var reader = new System.IO.StreamReader("FirstFile"))
{
//first we get columns for table 0
foreach (string s in reader.ReadLine().Split('\t'))
ds.Tables["File0"].Columns.Add(s);
while ((line = reader.ReadLine().Split('\t')) != null)
{
//and now the rest of the data.
var r = ds.Tables["File0"].NewRow();
for (int i = 0; i <= line.Length; i++)
{
r[i] = line[i];
}
ds.Tables["File0"].Rows.Add(r);
}
}
//we could probably do these in a loop or a second method,
//but you may want subtle differences, so for now we just do it the same way
//for file1
using (var reader2 = new System.IO.StreamReader("SecondFile"))
{
foreach (string s in reader2.ReadLine().Split('\t'))
ds.Tables["File1"].Columns.Add(s);
while ((line = reader2.ReadLine().Split('\t')) != null)
{
//and now the rest of the data.
var r = ds.Tables["File1"].NewRow();
for (int i = 0; i <= line.Length; i++)
{
r[i] = line[i];
}
ds.Tables["File1"].Rows.Add(r);
}
}
//you now have these in functioning datatables. Because we named columns,
//you can call them by name specifically, or by index, to replace in the first datatable.
string[] columnsToReplace = new string[] { "firstColumnName", "SecondColumnName", "ThirdColumnName" };
for(int i = 0; i < ds.Tables[0].Rows.Count; i++)
{
//you didn't give a sign of any relation between the two tables
//so this is just by row, and assumes the row count is equivalent.
//This is also not advised.
//if there is a key these sets of data share
//you should join on them instead.
foreach(DataRow dr in ds.Tables[0].Rows[i].ItemArray)
{
dr[3] = ds.Tables[1].Rows[i][columnsToReplace[0]];
dr[6] = ds.Tables[1].Rows[i][columnsToReplace[1]];
dr[11] = ds.Tables[1].Rows[i][columnsToReplace[2]];
}
}
//ds.Tables[0] now has the output you want.
string output = String.Empty;
foreach (var s in ds.Tables[0].Columns)
output = String.Concat(output, s ,"\t");
output = String.Concat(output, Environment.NewLine); // columns ready, now the rows.
foreach (DataRow r in ds.Tables[0].Rows)
output = string.Concat(output, r.ItemArray.SelectMany(t => (t.ToString() + "\t")), Environment.NewLine);
if(System.IO.File.Exists("MYPATH"))
using (System.IO.StreamWriter file = new System.IO.StreamWriter("MYPATH")) //or a variable instead of string literal
{
file.Write(output);
}
}
With Cinchoo ETL - an open source file helper library, you can do the merge of CSV files as below. Assumed the 2 CSV file contains same number of lines.
string CSV1 = #"Id Name City
1 Tom New York
2 Mark FairFax";
string CSV2 = #"Id City
1 Las Vegas
2 Dallas";
dynamic rec1 = null;
dynamic rec2 = null;
StringBuilder csv3 = new StringBuilder();
using (var csvOut = new ChoCSVWriter(new StringWriter(csv3))
.WithFirstLineHeader()
.WithDelimiter("\t")
)
{
using (var csv1 = new ChoCSVReader(new StringReader(CSV1))
.WithFirstLineHeader()
.WithDelimiter("\t")
)
{
using (var csv2 = new ChoCSVReader(new StringReader(CSV2))
.WithFirstLineHeader()
.WithDelimiter("\t")
)
{
while ((rec1 = csv1.Read()) != null && (rec2 = csv2.Read()) != null)
{
rec1.City = rec2.City;
csvOut.Write(rec1);
}
}
}
}
Console.WriteLine(csv3.ToString());
Hope it helps.
Disclaimer: I'm the author of this library.
Assume I have a .csv file with 70 columns, but only 5 of the columns are what I need. I want to be able to pass a method a string array of the columns names that I want, and for it to return a datatable.
private void method(object sender, EventArgs e) {
string[] columns =
{
#"Column21",
#"Column48"
};
DataTable myDataTable = Get_DT(columns);
}
public DataTable Get_DT(string[] columns) {
DataTable ret = new DataTable();
if (columns.Length > 0)
{
foreach (string column in columns)
{
ret.Columns.Add(column);
}
string[] csvlines = File.ReadAllLines(#"path to csv file");
csvlines = csvlines.Skip(1).ToArray(); //ignore the columns in the first line of the csv file
//this is where i need help... i want to use linq to read the fields
//of the each row with only the columns name given in the string[]
//named columns
}
return ret;
}
Read the first line of the file, line.Split(',') (or whatever your delimiter is), then get the index of each column name and store that.
Then for each other line, again do a var values = line.Split(','), then get the values from the columns.
Quick and dirty version:
string[] csvlines = File.ReadAllLines(#"path to csv file");
//select the indices of the columns we want
var cols = csvlines[0].Split(',').Select((val,i) => new { val, i }).Where(x => columns.Any(c => c == x.val)).Select(x => x.i).ToList();
//now go through the remaining lines
foreach (var line in csvlines.Skip(1))
{
var line_values = line.Split(',').ToList();
var dt_values = line_values.Where(x => cols.Contains(line_values.IndexOf(x)));
//now do something with the values you got for this row, add them to your datatable
}
You can look at https://joshclose.github.io/CsvHelper/
Think Reading individual fields is what you are looking for
var csv = new CsvReader( textReader );
while( csv.Read() )
{
var intField = csv.GetField<int>( 0 );
var stringField = csv.GetField<string>( 1 );
var boolField = csv.GetField<bool>( "HeaderName" );
}
We can easily do this without writing much code.
Exceldatareader is an awesome dll for that, it will directly as a datable from the excel sheet with just one method.
here is the links for example:http://www.c-sharpcorner.com/blogs/using-iexceldatareader1
http://exceldatareader.codeplex.com/
Hope it was useful kindly let me know your thoughts or feedbacks
Thanks
Karthik
var data = File.ReadAllLines(#"path to csv file");
// the expenses row
var query = data.Single(d => d[0] == "Expenses");
//third column
int column21 = 3;
return query[column21];
As others have stated a library like CsvReader can be used for this. As for linq, I don't think its suitable for this kind of job.
I haven't tested this but it should get you through
using (TextReader textReader = new StreamReader(filePath))
{
using (var csvReader = new CsvReader(textReader))
{
var headers = csvReader.FieldHeaders;
for (int rowIndex = 0; csvReader.Read(); rowIndex++)
{
var dataRow = dataTable.NewRow();
for (int chosenColumnIndex = 0; chosenColumnIndex < columns.Count(); chosenColumnIndex++)
{
for (int headerIndex = 0; headerIndex < headers.Length; headerIndex++)
{
if (headers[headerIndex] == columns[chosenColumnIndex])
{
dataRow[chosenColumnIndex] = csvReader.GetField<string>(headerIndex);
}
}
}
dataTable.Rows.InsertAt(dataRow, rowIndex);
}
}
}
I am building a tool to automate the creation of an Excel workbook that contains a table and an associated PivotTable. The table structure is on one sheet, the data for which will be pulled from a database using another tool at a later point. The PivotTable is on a second sheet using the table from the previous sheet as the source.
I am using EPPlus to facilitate building the tool but am running into problems specifying the cacheSource. I am using the following to create the range and PivotTable:
var dataRange = dataWorksheet.Cells[dataWorksheet.Dimension.Address.ToString()];
var pivotTable = pivotWorksheet.PivotTables.Add(pivotWorksheet.Cells["B3"], dataRange, name);
This sets the cacheSource to:
<x:cacheSource type="worksheet" xmlns:x="http://schemas.openxmlformats.org/spreadsheetml/2006/main">
<x:worksheetSource ref="A1:X2" sheet="dataWorksheet" />
or within Excel, the data source is set to:
dataWorksheet!$A$1:$X$2
This works fine if the table size never changes, but as the number of rows will be dynamic, I am finding when the data is refreshed, data is only read from the initial range specified.
What I am want to do is to programmatically set the cacheSource to:
<x:cacheSource type="worksheet" xmlns:x="http://schemas.openxmlformats.org/spreadsheetml/2006/main">
<x:worksheetSource name="dataWorksheet" />
</x:cacheSource>
or in Excel, set the data source to:
dataWorksheet
I believe it may be possible to do this by accessing the XML directly (any pointers on this would be most welcome) but is there any way to do this using EPPlus?
It can be done but it is not the prettiest thing in the world. You can extract the cache def xml and edit it from the created EPPlus pivot table object but that will wreak havoc with the save logic when you call package.save() (or GetAsByteArray()) since it parses the xml on save to generate the final file. This is the result of, as you said, EPPlus not capable of handling a table as the source.
So, your alternative is to save the file with EPPlus normally and then manipulate the content of the xlsx which is a renamed zip file using a .net ZipArchive. The trick is you cannot manipulate the files out of order in the zip otherwise Excel will complain when it opens the file. And since you cannot insert an entry (only add to the end) you have to recreate the zip. Here is an extension method on a ZipArchive that will allow you to update the cache source:
public static bool SetCacheSourceToTable(this ZipArchive xlsxZip, FileInfo destinationFileInfo, string tablename, int cacheSourceNumber = 1)
{
var cacheFound = false;
var cacheName = String.Format("pivotCacheDefinition{0}.xml", cacheSourceNumber);
using (var copiedzip = new ZipArchive(destinationFileInfo.Open(FileMode.Create, FileAccess.ReadWrite), ZipArchiveMode.Update))
{
//Go though each file in the zip one by one and copy over to the new file - entries need to be in order
xlsxZip.Entries.ToList().ForEach(entry =>
{
var newentry = copiedzip.CreateEntry(entry.FullName);
var newstream = newentry.Open();
var orgstream = entry.Open();
//Copy all other files except the cache def we are after
if (entry.Name != cacheName)
{
orgstream.CopyTo(newstream);
}
else
{
cacheFound = true;
//Load the xml document to manipulate
var xdoc = new XmlDocument();
xdoc.Load(orgstream);
//Get reference to the worksheet xml for proper namespace
var nsm = new XmlNamespaceManager(xdoc.NameTable);
nsm.AddNamespace("default", xdoc.DocumentElement.NamespaceURI);
//get the source
var worksheetSource = xdoc.SelectSingleNode("/default:pivotCacheDefinition/default:cacheSource/default:worksheetSource", nsm);
//Clear the attributes
var att = worksheetSource.Attributes["ref"];
worksheetSource.Attributes.Remove(att);
att = worksheetSource.Attributes["sheet"];
worksheetSource.Attributes.Remove(att);
//Create the new attribute for table
att = xdoc.CreateAttribute("name");
att.Value = tablename;
worksheetSource.Attributes.Append(att);
xdoc.Save(newstream);
}
orgstream.Close();
newstream.Flush();
newstream.Close();
});
}
return cacheFound;
}
And here is how to use it:
//Throw in some data
var datatable = new DataTable("tblData");
datatable.Columns.AddRange(new[]
{
new DataColumn("Col1", typeof (int)), new DataColumn("Col2", typeof (int)), new DataColumn("Col3", typeof (object))
});
for (var i = 0; i < 10; i++)
{
var row = datatable.NewRow();
row[0] = i; row[1] = i*10; row[2] = Path.GetRandomFileName();
datatable.Rows.Add(row);
}
const string tablename = "PivotTableSource";
using (var pck = new ExcelPackage())
{
var workbook = pck.Workbook;
var source = workbook.Worksheets.Add("source");
source.Cells.LoadFromDataTable(datatable, true);
var datacells = source.Cells["A1:C11"];
source.Tables.Add(datacells, tablename);
var pivotsheet = workbook.Worksheets.Add("pivot");
pivotsheet.PivotTables.Add(pivotsheet.Cells["A1"], datacells, "PivotTable1");
using (var orginalzip = new ZipArchive(new MemoryStream(pck.GetAsByteArray()), ZipArchiveMode.Read))
{
var fi = new FileInfo(#"c:\temp\Pivot_From_Table.xlsx");
if (fi.Exists)
fi.Delete();
var result = orginalzip.SetCacheSourceToTable(fi, tablename, 1);
Console.Write("Cache source was updated: ");
Console.Write(result);
}
}