as you can see, the package starts by reading from an excel sheet, then convert the columns to another types, then the result of the conversion is the input on a lookup transformation. one of the columns that the data conversion changes type is named "badge". in the excel sheet this "badge" column has numeric data type. the data conversion convert the "badge" from numeric to "Unicode string [DT_WSTR]" (also i tried "string [DT_STR]"). the lookup should compare between the input from the data conversion and a table in a database that contains a column also named "badge" with a data type of nvarchar(50). when i try to do maping on the "badge" the following error appears: Cannot map the lookup column, 'badge', because the column is set to a floating point data type.

"Just an insanity check. you're mapping the new column the data converted not the original one" let's break down. first double click on the green arrow between the conversion and the lookup. and in the Metadata make sure that the data type is converted right. then you may try to replace the data conversion with the Drived Column and add new column (DT_WSTR)badge. I know this is not much of help but this what i would doWagdy Ishac www.sqldair.com

thanks, you're right ... the data type in the Metadata is still DT_R8 !! now the questions are, what is the use of the data conversion if it didn't convert ?! and what is the solution ? thanks again

ok ... if you scroll down in the Path Metadata list in the Data Flow Path Editor and the Avilable Input Columns in the loockup transformation editor, you will find additional columns starts with "Copy of " then a column name (in my case "Copy of badge"). those additional columns have the desired and new data types. you can change the aliases of those new column in the data conversion transformation editor through the Output Alias column.