If both of your tables are in SQL Server, just issue the relevant INSERT and UPDATE queries using Execute SQL tasks ...

Or avoid SSIS completely and use a standard SQL...

It's always easier to find something when you know roughly where it is:-D Glad I could help.

On your dataflow, if your source is DataReader, Excel or OLE DB, you have the option of setting the SQL Command property within that source.

In your case, set it along...

Not done it myself, but in the Execute SQL task change source type to FileConnection and then set the FileConnection property accordingly ... Looks like it should work.

I made this work by creating a script transformation component and adding an extra 'RowNum' column as an output.

Within the script component, after defining the new output column, you just...

See my preceding post - the line in bold. It was not showing correctly before, so I just edited it. Maybe that was your problem.

Add a Script task after your SQL task.

Click on Script and In ReadOnlyVariables, enter User::InDatabase

Click on Design Script and enter just one line in Main() to make it look like...

Edit the task. Choose 'Expressions' and create a new one by clicking on the ellipsis, then select SourceDatabaseName as your property and enter the variable expression under 'Expression' ...

I think I would handle this slightly differently.

I would split the input columns so that the numeric data is in one column and the 'special' character is in another.

I would...

There is an SSIS Data Flow Transformation called a Merge Join, which takes as its inputs (sorted) data from two datasets and merges them in the way you want.

The staging...

Or, depending on the nature of your source, you could have used 'SQL Command' as your data access mode and added a suitable WHERE clause to exclude the unwanted data.

That is not what I mean. If you right click white space in the control flow window and select Variables, a list will be displayed. Column 1 is variable name,...

Can you confirm the scope of your variables?

Unless this needs to be done in (near) real time, a scheduled package containing a simple foreach loop should be sufficient for your needs ...

I added a derived column called 'Match' to a test transformation - expression as follows:

price == floor(price)

where price is a numeric(17,4) field and it worked perfectly, using the examples you...