asked on

Efficient Datatable value manipulation

Hello Experts,

I'm developing an application in Visual Studio 2008 to parse and coalesce multiple laboratory results from *.csv into one single DataTable. Resultant tables tend to contain 1000+ records over >30 columns. The user is intended to first view the data in a DataGrid and then manipulate select columns accordingly using a set of functions (generally removing erroneous characters such as "<", ">"). The data is then refreshed to the DataGrid.

My current method of is slow and inefficient, taking ~45 seconds per column, per 1000 records. Is there a better way to manipulate data within datatables? I realise manipulating the underlying data prior to visualisation would be better, however the user must see the data first to choose effective treatments.

Any help would be appreciated.

Olly

Function ConvertNo(ByVal Factor As Decimal, ByVal colNam As String, ByVal chrX As String)
 
Dim inVal = Nothing
 
        For Each dRow As DataRow In tab_Data.Rows
            inVal = dRow(colNam)
            If Not IsDBNull(inVal) Then
                inVal = Val(Regex.Replace(inVal, chrX, "")) * Factor
                dRow(colNam) = inVal
            End If
        Next dRow
 
        Return inVal
 
End Function

Open in new window

SOLUTION

MasterPage

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Groovyolly

ASKER

Cheers for that.

That speeds things up a little, however, when multiple columns are selected the processing time is still considerable.

I'm curious if there's a different approach to the standard compare - replace loop available to DataTables. Something more along the lines of a Find - Replace which would be more efficient.

Anurag Thakur

first thing i noticed is that you are looping for all the rows to update just one column whether there are invalid characters or not
what you should be doing is find all the rows first for a column where there are invalid characters

DataRow [] rows = dt.Select ("ColumnName like '%" + strOne + "'% or ColumnName Like '% + strTwo + "%'");
if (rows != null)
{
for (int i = 0; i < rows.Lenght; i++)
{
// update that column value
}
}

dt.AcceptChanges();

Groovyolly

ASKER

This is a better solution fundamentally, however I'm having some trouble applying it to my situation.

I'm getting a 'Missing operand after 'Mod' operator' error with the following code:

Dim drSelect As DataRow() = tab_Data.Select("ColumnName like '%" + colNam + "'%")

Open in new window

ASKER CERTIFIED SOLUTION

Anurag Thakur

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Groovyolly

ASKER

Sadly this returns a 'Cannot find column [ColumnName]'. I've tried switching the expression around a little, with no effect.

Anurag Thakur

Does the datatable contain a colulm named COLUMNNAME -
'Cannot find column [ColumnName]'

you need to replace the name COLUMNNAME with the name of the column in the datatable on which you want to do the operations

Groovyolly

ASKER

No column called 'ColumnName' in there - the datatable columns are all added according to the parsed input.

With respect to replacing ColumnName with my column value, in the statement tab_Data.Select("ColumnName like '%" + colNam + "%'") colName represents the column name passed into the function, so (using a column called "A", for example) the function effectively reads tab_Data.Select("ColumnName Like '%A%'"). To replace ColumnName directly would result in ("A Like '%A%'"), which doesn't make sense.

Using a watch applied to the bracketed portion of the statement, I've determined that the statement is reading ("ColumnName Like '%A%'"). I'm wondering whether ColumnName is a valid reference to the column title in the Select function as it doesn't appear to be accepted by the function.

Anurag Thakur

even if you pass the column name and the value you want to compare to a function
you can use my solution like this
-- i am assuming that you know the column name in the datatable on which the operations are to be done

private void ReplaceSomething(string columnName, string ColumnValue)
{
DataRows [] rows = tab_Data.Select(columnName + " like '%" + ColumnValue + "%'")
}

Groovyolly

ASKER

Seems the problem was that my filter string contains a hyphen special character, so colNam needed to be surrounded by square brackets. The column name was being truncated at the hyphen, causing an error (despite this, the string actually looked fine when watched).

For reference I've included the corrected code.

I've used MasterPage's suggestion in the replace string too.

Thanks to both of you for your time and patience.

Dim drSelect As DataRow() = tab_Data.Select("[" & colNam & "] Like '%" + chrX + "%'")

Open in new window