I'm trying to port some old VB6 code to C# and .NET.
There are a number of places where the old code uses a RecordSet to execute a SQL query and then loop through the results. No problem so far, but inside the loop the code makes changes to the current row, updating columns and even deleting the current row altogether.
In .NET, I can easily use a SqlDataReader to loop through SQL query results, but updates are not supported.
So I've been playing with using a SqlDataAdapter to populate a DataSet, and then loop through the rows in a DataSet table. But the DataSet doesn't seem very smart compared to the VB6's old RecordSet. For one thing, I need to provide update queries for each type of edit I have. Another concern is that a DataSet seems to hold everything in memory at once, which might be a problem if there are many results.
What is the best way to duplicate this behavior in .NET? The code below shows what I have so far. Is this the best approach, or is there another option?
using (SqlConnection connection = new SqlConnection(connectionString))
{
DataSet dataset = new DataSet();
using (SqlDataAdapter adapter = new SqlDataAdapter(new SqlCommand(query, connection)))
{
adapter.Fill(dataset);
DataTable table = dataset.Tables[0];
foreach (DataRow row in table.Rows)
{
if ((int)row["Id"] == 4)
{
if ((int)row["Value1"] > 0)
row["Value2"] = 12345;
else
row["Value3"] = 12345;
}
else if ((int)row["Id"] == 5)
{
row.Delete();
}
}
// TODO:
adapter.UpdateCommand = new SqlCommand("?", connection);
adapter.DeleteCommand = new SqlCommand("?", connection);
adapter.Update(table);
}
}
Note: I'm new to the company and can't very well tell them they have to change their connection strings or must switch to Entity Framework, which would be my choice. I'm really looking for a code-only solution.
ADO.NET DataTable and DataAdapter provide the closest equivalent of ADO Recordset with applies separation of concens principle. DataTable contains the data and provides the change tracking information (similar to EF internal entity tracking) while DataAdapter provides a standard way to populate it from database (Fill method) and apply changes back to the database (Update method).
With that being said, what are you doing is the intended way to port the ADO Recordset to ADO.NET. The only thing you've missed is that you are not always required to specify Insert, Update and Delete commands. As soon as your query is querying a single table (which I think was a requirement to get updateable Recordset anyway), you can use another ADO.NET player called DbCommandBuilder:
Automatically generates single-table commands used to reconcile changes made to a DataSet with the associated database.
Every database provider provides implementation of this abstract class. The MSDN example for SqlCommandBuilder is almost identical to your sample, so all you need before calling Update is (a bit counterintuitive):
var builder = new SqlCommandBuilder(adapter);
and that's it.
Behind the scenes,
The DbCommandBuilder registers itself as a listener for RowUpdating events that are generated by the DbDataAdapter specified in this property.
and dynamically generates the commands if they are not specifically set in the data adapter by you.
I came up with an (untested) solution for a data table.
It does require you to do some work, but it should generate update and delete commands for each row you change or delete automatically, by hooking up to the RowChanged and RowDeleted events of the DataTable.
Each row will get it's own command, equivalent to ADODB.RecordSet update / delete methods.
However, unlike the ADODB.RecordSet methods, this class will not change the underling database, but only create the SqlCommands to do it. Of course, you can change it to simply execute them on once they are created, but as I said, I didn't test it so I'll leave that up to you if you want to do it. However, please note I'm not sure how the RowChanged event will behave for multiple changes to the same row. Worst case it will be fired for each change in the row.
The class constructor takes three arguments:
The instance of the DataTable class you are working with.
A Dictionary<string, SqlDbType> that provides mapping between column names and SqlDataTypes
An optional string to represent table name. If omitted, the TableName property of the DataTable will be used.
Once you have the mapping dictionary, all you have to do is instantiate the CommandGenerator class and iterate the rows in the data table just like in the question. From that point forward everything is automated.
Once you completed your iteration, all you have to do is get the sql commands from the Commands property, and run them.
public class CommandGenerator
{
private Dictionary<string, SqlDbType> _columnToDbType;
private string _tableName;
private List<SqlCommand> _commands;
public CommandGenerator(DataTable table, Dictionary<string, SqlDbType> columnToDbType, string tableName = null)
{
_commands = new List<SqlCommand>();
_columnToDbType = columnToDbType;
_tableName = (string.IsNullOrEmpty(tableName)) ? tableName : table.TableName;
table.RowDeleted += table_RowDeleted;
table.RowChanged += table_RowChanged;
}
public IEnumerable<SqlCommand> Commands { get { return _commands; } }
private void table_RowChanged(object sender, DataRowChangeEventArgs e)
{
_commands.Add(GenerateDelete(e.Row));
}
private void table_RowDeleted(object sender, DataRowChangeEventArgs e)
{
_commands.Add(GenerateDelete(e.Row));
}
private SqlCommand GenerateUpdate(DataRow row)
{
var table = row.Table;
var cmd = new SqlCommand();
var sb = new StringBuilder();
sb.Append("UPDATE ").Append(_tableName).Append(" SET ");
var valueColumns = table.Columns.OfType<DataColumn>().Where(c => !table.PrimaryKey.Contains(c));
AppendColumns(cmd, sb, valueColumns, row);
sb.Append(" WHERE ");
AppendColumns(cmd, sb, table.PrimaryKey, row);
cmd.CommandText = sb.ToString();
return cmd;
}
private SqlCommand GenerateDelete(DataRow row)
{
var table = row.Table;
var cmd = new SqlCommand();
var sb = new StringBuilder();
sb.Append("DELETE FROM ").Append(_tableName).Append(" WHERE ");
AppendColumns(cmd, sb, table.PrimaryKey, row);
cmd.CommandText = sb.ToString();
return cmd;
}
private void AppendColumns(SqlCommand cmd, StringBuilder sb, IEnumerable<DataColumn> columns, DataRow row)
{
foreach (var column in columns)
{
sb.Append(column.ColumnName).Append(" = #").AppendLine(column.ColumnName);
cmd.Parameters.Add("#" + column.ColumnName, _columnToDbType[column.ColumnName]).Value = row[column];
}
}
}
As I wrote, this is completely untested, but I think it should be enough to at least show the general idea.
Your constraints:
Not using Entity Framework
DataSet seems to hold everything in memory at once, which might be a
problem if there are many results.
a code-only solution ( no external libraries)
Plus
The maximum number of rows that a DataTable can store is 16,777,216
row MSDN
To get high performance
//the main class to update/delete sql batches without using DataSet/DataTable.
public class SqlBatchUpdate
{
string ConnectionString { get; set; }
public SqlBatchUpdate(string connstring)
{
ConnectionString = connstring;
}
public int RunSql(string sql)
{
using (SqlConnection con = new SqlConnection(ConnectionString))
using (SqlCommand cmd = new SqlCommand(sql, con))
{
cmd.CommandType = CommandType.Text;
con.Open();
int rowsAffected = cmd.ExecuteNonQuery();
return rowsAffected;
}
}
}
//------------------------
// using the class to run a predefined patches
public class SqlBatchUpdateDemo
{
private string connstring = "myconnstring";
//run batches in sequence
public void RunBatchesInSequence()
{
var sqlBatchUpdate = new SqlBatchUpdate(connstring);
//batch1
var sql1 = #"update mytable set value2 =1234 where id =4 and Value1>0;";
var nrows = sqlBatchUpdate.RunSql(sql1);
Console.WriteLine("batch1: {0}", nrows);
//batch2
var sql2 = #"update mytable set value3 =1234 where id =4 and Value1 =0";
nrows = sqlBatchUpdate.RunSql(sql2);
Console.WriteLine("batch2: {0}", nrows);
//batch3
var sql3 = #"delete from mytable where id =5;";
nrows = sqlBatchUpdate.RunSql(sql3);
Console.WriteLine("batch3: {0}", nrows);
}
// Alternative: you can run all batches as one
public void RunAllBatches()
{
var sqlBatchUpdate = new SqlBatchUpdate(connstring );
StringBuilder sb = new StringBuilder();
var sql1 = #"update mytable set value2 =1234 where id =4 and Value1>0;";
sb.AppendLine(sql1);
//batch2
var sql2 = #"update mytable set value3 =1234 where id =4 and Value1 =0";
sb.AppendLine(sql2);
//batch3
var sql3 = #"delete from mytable where id =5;";
sb.AppendLine(sql3);
//run all batches
var nrows = c.RunSql(sb.ToString());
Console.WriteLine("all patches: {0}", nrows);
}
}
I simulated that solution and it's working fine with a high performance because all updates /delete run as batch.
Related
I'm loading data into a form with 3 Entry controls.
The object I am using for this is called mySettings, which is an object of SystemSettings, a class and database table in my SQLite database.
So far I have this code, and it works as is.
var db = new SQLiteConnection(dbPath);
Entry txtServer;
txtServer = new Entry { FontSize = 10 };
controlGrid.Children.Add(txtServer, 2, 0);
Grid.SetColumnSpan(txtServer, 4);
SystemSettings mySettings;
mySettings = db.Get<SystemSettings>(0);
txtServer.Text = mySettings.FTPServer;
However, I need to check whether SystemSettings contains any rows in the table before I load values in.
I've seen a few guides online.
Some say use something along the lines of
SQLiteCommand cmd;
cmd = new SQLiteCommand(db);
...
int result = Convert.ToInt32(db.ExecuteScalar)
However, I get an error there saying
SQLiteCommand does not contain any method containing x parameters
no matter how many I pass in (0 or more).
There also doesn't appear to be a method as part of db.
So how can I check whether SystemSettings contains any rows, before trying to use data that doesn't exist?
The pattern below should work. The .ExecuteScalar() method is actually on the command and not the connection.
int count;
using (SQLiteConnection db = new SQLiteConnection("MY_CXN_STRING"))
using (SQLiteCommand cmd = new SQLiteCommand("SELECT COUNT(*) FROM SystemSettings"))
{
db.Open();
count = (int)cmd.ExecuteScalar();
db.Close();
}
bool hasRows = count != 0;
Basically you want to clear
SystemSettings
Try just running a query that returns nothing against the database. For instance:
SystemSettings = $"SELECT * FROM TABLE_NAME WHERE COLUMN_NAME IS 'INVALID_EXPRESSIONdjeiq48724rufnjdrandom stuff'";
Not the most elegant solution by any means, but it works.
What you want to do is to get the first row in you SystemSettings table if any:
You should therefore execute the following Sql Statement (or something similar) and check if a result is returned:
Select * from SystemSettings LIMIT 1;
You can execute the query and check the result like this:
public bool DoesTableContainRows(string tableName, SQLiteConnection connection)
{
var command = new SQLiteCommand($"Select * from {tableName } LIMIT 1;", connection);
var resultReader = command.ExecuteReader();
// check whether or not a row was returned
bool containRows = resultReader.Read();
resultReader.Close();
return containRows;
}
Edit:
Shows how to check if a table contains rows using .NET and Microsoft.Data.Sqlite including better disposing of resources.
public bool DoesTableContainRows(string tableName, SqliteConnection connection)
{
using (var command = new SqliteCommand($"Select * from {tableName } LIMIT 1;", connection))
{
using (var resultReader = command.ExecuteReader())
{
// check whether or not a row was returned
bool containRows = resultReader.Read();
resultReader.Close();
return containRows;
}
}
}
I am trying to write a Method that simply accepts a string which is an SQL Command and runs it against the pre-defined Database/Server ... This is what I have so far:
public class TSqlConnector : IDbConnector
{
private readonly string _connectionString;
public TSqlConnector(string conn)
{
_connectionString = conn;
}
public IEnumerable<object[]> ExecuteCommand(string query)
{
var res = new List<object[]>();
try
{
using (SqlConnection sql = new SqlConnection(_connectionString))
{
sql.Open();
SqlCommand cmd = new SqlCommand(query, sql);
var reader = cmd.ExecuteReader();
DataTable tbl = new DataTable();
while (reader.Read())
{
var dr = tbl.NewRow();
dr.ItemArray = new object[reader.FieldCount];
reader.GetValues(dr.ItemArray);
res.Add(dr.ItemArray);
}
}
return res;
}
catch (Exception ex)
{
Console.WriteLine(ex.ToString());
throw;
}
}
}
This code, however, gives me an error saying that
Input array is longer than the number of columns in this table.
I googled the error message, apparently I first have to define the DataTable's columns, using tbl.Add("ColumnName", typeof(type));
This however, completely undermines what I was trying to do - writing a generic version. All I wanted was some kind of construct which contains the information I would get from the SqlServer if I typed the same command into SSMS, I don't really care what hoops I have to jump through to read the data in C#; Using an object-Array for each row, having to manually cast each object into a string, int or whatever is perfectly acceptable, even a CSV-like string would be just fine
The only thing I don't want to do is add a definition for the table or a fixed amount of row. Each method that uses ExecuteCommand() will have to know what type of object-array is returned and that's fine but adding some complex data structure containing types and column names in addition to the SQL Commands seems like overkill.
Is there some easier way to achieve this?
What you have is an IDataReader from your cmd.ExecuteReader();
To load the results into a DataTable, you can use the Load method as follows:
var reader = cmd.ExecuteReader();
DataTable tbl = new DataTable();
tbl.Load(reader);
// now tbl contains the corresponding columns and rows from your sql command.
// Then you can return the ItemArrays from each row;
return tbl.Rows.Cast<DataRow>().Select(row => row.ItemArray);
I've used code like this to input a generic SQL query and return the results as a datatable. Of course, you'll have to parse out the results how you need them.
private DataTable QueryToTable(string sql, string cs)
{
var ds = new DataSet();
using (var adapter = new SqlDataAdapter(sql, cs))
{
adapter.Fill(ds);
}
return ds.Tables(0);
}
I'm trying to do some manual SQL queries against my SQLite database using the ExecuteStoreQuery method on my ObjectContext.
The catch is that I don't always know how many columns are in the table that I'm querying. Ideally, I would like each fetched row to simply be an string[] object.
I've looked at Example 2 here: http://msdn.microsoft.com/en-us/library/vstudio/dd487208(v=vs.100).aspx
It's close to what I want to do, except that I don't know the structure of the TElement I'm fetching, so I can't define a struct as they do in the example.
Below is some of my code (not compiling due to the ???? TElement). The code below is trying to fetch the table info, so in this case I do know the structure of the rows, but in general I don't.
Is there a way to do this with ExecuteStoreQuery? Or is there a different way of doing it, while still using the existing connection of my ObjectContext (rather than opening a new SQL connection to the DB)?
public void PrintColumnHeaders(NWRevalDatabaseEntities entities, string tableName)
{
string columnListQuery = string.Format("PRAGMA table_info({0})", tableName);
var result = entities.ExecuteStoreQuery<????>(columnListQuery);
foreach (string[] row in result)
{
string columnHeader = row[1]; // Column header is in second column of table
Console.WriteLine("Column Header: {0}", columnHeader);
}
}
I got this working based on Gert Arnold's comment. Also, it took me some effort to figure out that I need a SQLiteConnection, not the EntityConnection that I could get directly from the ObjectContext. The answer to this question helped me with that.
The working code is below:
public static void PrintColumnHeaders(NWRevalDatabaseEntities entities, string tableName)
{
var sc = ((System.Data.EntityClient.EntityConnection)entities.Connection).StoreConnection;
System.Data.SQLite.SQLiteConnection sqliteConnection = (System.Data.SQLite.SQLiteConnection)sc;
sqliteConnection.Open();
System.Data.Common.DbCommand cmd = sc.CreateCommand();
cmd.CommandType = System.Data.CommandType.Text;
cmd.CommandText = string.Format("PRAGMA table_info('{0}');", tableName);
System.Data.Common.DbDataReader reader = cmd.ExecuteReader();
if (reader.HasRows)
{
object[] values = new object[reader.FieldCount];
while (reader.Read())
{
int result = reader.GetValues(values);
string columnHeader = (string)values[1]; // table_info returns a row for each column, with the column header in the second column.
Console.WriteLine("Column Header: {0}", columnHeader);
}
}
sqliteConnection.Close();
}
I have an asp.net project done in C# (my C# syntax is very rusty) and am using the built in database it creates with the project. I've created a table called aspnet_Tutorials that (for now) stores two columns of user submitted data: TutorialName and TutorialContent. Very simple, it's a learning project.
What I need to do is create a list from the first column of aspnet_Tutorials to use it to create a "directory" of the tutorials on a page. The part I'm having trouble with, mostly syntactically, is connecting to and iterating over the column to get the values into a list. Could anyone provide a straight forward example of this? And possibly explain what's going on in the code.
public class TutorialsDirDAL
{
SqlConnection conn = new SqlConnection(ConfigurationManager.ConnectionStrings["ApplicationServices"].ConnectionString);
SqlCommand cmd = new SqlCommand();
SqlDataAdapter da = new SqlDataAdapter();
DataSet ds = new DataSet();
public List<string> DisplayTutorials() //parameters? String qry?
{
//query the database table, foreach loop over the data, place it into a list?
}
}
I know how to write simple sql queries. But I've seen a couple different set ups for this in my Googling spree. I currently have the following for a query, feel free to pick it apart or offer a better solution than using a query.
cmd.CommandText = "SELECT * FROM ASPNET_TUTORIALS (TutorialTitle)"
+ "VALUES (#tutorialTitle)";
Thank you!
ebad86's answer is acceptable but since you are obviously learning I think introducing OO principals muddy the water with what you are trying to learn at this point.
Here is a basic method:
private void GetData()
{
//The object that will physically connect to the database
using(SqlConnection cnx = new SqlConnection("<your connection string>")
{
//The SQL you want to execute
SqlCommand cmd = new SqlCommand("SELECT * FROM ASPNET_TUTORIALS");
//Open the connection to the database
cnx.Open();
//execute your command
using (IDataReader dataReader = cnx.ExecuteReader(cmd))
{
//Loop through your results
while(dataReader.Read())
{
//do whatever to the data
ListItem item = new ListItem(Convert.ToString(dataReader["TutorialName"]));
lst.Items.Add(item);
}
}
}
}
This is all very straightforward. The part you are most interested in is the while loop though. The loop will go through all of the returned records and you can do whatever you need to do with them. In my example I have assumed that there is a ListBox named 'lst' and I am simply adding ListItems to it that will have the name of whatever 'TutorialName' is. You can make literally do whatever you need to do with the data at this point. To fit your example (returning a List) you would do this:
private List<string> GetData()
{
List<string> lst = new List<string>();
//The object that will physically connect to the database
using(SqlConnection cnx = new SqlConnection("<your connection string>")
{
//The SQL you want to execute
SqlCommand cmd = new SqlCommand("SELECT * FROM ASPNET_TUTORIALS");
//Open the connection to the database
cnx.Open();
//execute your command
using (IDataReader dataReader = cnx.ExecuteReader(cmd))
{
//Loop through your results
while(dataReader.Read())
{
lst.Add(Convert.ToString(dataReader["TutorialName"]));
}
}
}
return lst;
}
Please respond if you have any questions.
Well you can fetch using data reader and map to the object. I can give you some rough code, which could be like this:
using (IDataReader objDataReader = objDB.ExecuteReader(objCMD))
{
while (objDataReader.Read())
{
DataBaseObject obj = new DataBaseObject();
obj = MapObjectToList(objDataReader);
ObjectList.Add(obj);
}
objDataReader.Dispose();
}
// Mapping Function can be called somewhat like this:
private DataBaseObject MapObjectToList(IDataReader objDataReader)
{
DataBaseObject obj = new DataBaseObject();
obj.Prop1Name = base.GetDataValue<string>(objDataReader, "Column1Name");
obj.Prop2Name = base.GetDataValue<string>(objDataReader, "Column2Name");
return obj;
}
But this is just a rough idea how I would do it.
I have a DataTable that is filled in from an SQL query to a local database, but I don't know how to extract data from it.
Main method (in test program):
static void Main(string[] args)
{
const string connectionString = "server=localhost\\SQLExpress;database=master;integrated Security=SSPI;";
DataTable table = new DataTable("allPrograms");
using (var conn = new SqlConnection(connectionString))
{
Console.WriteLine("connection created successfuly");
string command = "SELECT * FROM Programs";
using (var cmd = new SqlCommand(command, conn))
{
Console.WriteLine("command created successfuly");
SqlDataAdapter adapt = new SqlDataAdapter(cmd);
conn.Open();
Console.WriteLine("connection opened successfuly");
adapt.Fill(table);
conn.Close();
Console.WriteLine("connection closed successfuly");
}
}
Console.Read();
}
The command I used to create the tables in my database:
create table programs
(
progid int primary key identity(1,1),
name nvarchar(255),
description nvarchar(500),
iconFile nvarchar(255),
installScript nvarchar(255)
)
How can I extract data from the DataTable into a form meaningful to use?
The DataTable has a collection .Rows of DataRow elements.
Each DataRow corresponds to one row in your database, and contains a collection of columns.
In order to access a single value, do something like this:
foreach(DataRow row in YourDataTable.Rows)
{
string name = row["name"].ToString();
string description = row["description"].ToString();
string icoFileName = row["iconFile"].ToString();
string installScript = row["installScript"].ToString();
}
You can set the datatable as a datasource to many elements.
For eg
gridView
repeater
datalist
etc etc
If you need to extract data from each row then you can use
table.rows[rowindex][columnindex]
or
if you know the column name
table.rows[rowindex][columnname]
If you need to iterate the table then you can either use a for loop or a foreach loop like
for ( int i = 0; i < table.rows.length; i ++ )
{
string name = table.rows[i]["columnname"].ToString();
}
foreach ( DataRow dr in table.Rows )
{
string name = dr["columnname"].ToString();
}
The simplest way to extract data from a DataTable when you have multiple data types (not just strings) is to use the Field<T> extension method available in the System.Data.DataSetExtensions assembly.
var id = row.Field<int>("ID"); // extract and parse int
var name = row.Field<string>("Name"); // extract string
From MSDN, the Field<T> method:
Provides strongly-typed access to each of the column values in the
DataRow.
This means that when you specify the type it will validate and unbox the object.
For example:
// iterate over the rows of the datatable
foreach (var row in table.AsEnumerable()) // AsEnumerable() returns IEnumerable<DataRow>
{
var id = row.Field<int>("ID"); // int
var name = row.Field<string>("Name"); // string
var orderValue = row.Field<decimal>("OrderValue"); // decimal
var interestRate = row.Field<double>("InterestRate"); // double
var isActive = row.Field<bool>("Active"); // bool
var orderDate = row.Field<DateTime>("OrderDate"); // DateTime
}
It also supports nullable types:
DateTime? date = row.Field<DateTime?>("DateColumn");
This can simplify extracting data from DataTable as it removes the need to explicitly convert or parse the object into the correct types.
Please consider using some code like this:
SqlDataReader reader = command.ExecuteReader();
int numRows = 0;
DataTable dt = new DataTable();
dt.Load(reader);
numRows = dt.Rows.Count;
string attended_type = "";
for (int index = 0; index < numRows; index++)
{
attended_type = dt.Rows[indice2]["columnname"].ToString();
}
reader.Close();
Unless you have a specific reason to do raw ado.net I would have a look at using an ORM (object relational mapper) like nHibernate or LINQ to SQL. That way you can query the database and retrieve objects to work with which are strongly typed and easier to work with IMHO.
var table = Tables[0]; //get first table from Dataset
foreach (DataRow row in table.Rows)
{
foreach (var item in row.ItemArray)
{
console.Write("Value:"+item);
}
}
Please, note that Open and Close the connection is not necessary when using DataAdapter.
So I suggest please update this code and remove the open and close of the connection:
SqlDataAdapter adapt = new SqlDataAdapter(cmd);
conn.Open(); // this line of code is uncessessary
Console.WriteLine("connection opened successfuly");
adapt.Fill(table);
conn.Close(); // this line of code is uncessessary
Console.WriteLine("connection closed successfuly");
Reference Documentation
The code shown in this example does not explicitly open and close the
Connection. The Fill method implicitly opens the Connection that the
DataAdapter is using if it finds that the connection is not already
open. If Fill opened the connection, it also closes the connection
when Fill is finished. This can simplify your code when you deal with
a single operation such as a Fill or an Update. However, if you are
performing multiple operations that require an open connection, you
can improve the performance of your application by explicitly calling
the Open method of the Connection, performing the operations against
the data source, and then calling the Close method of the Connection.
You should try to keep connections to the data source open as briefly
as possible to free resources for use by other client applications.