I have an Excel data source with a column named policy_ref which contains data in a mixture of formats. Mostly the format of policy_ref is along the lines of ABC/123456. It so happens that the policy_ref values at the top of the Excel sheet contain refs in the format 123456, which Excel has automatically detected as numeric. I want to import this Excel sheet into a SQL table using a SSIS data flow task, converting policy_ref into varchar(255). To achieve this I have placed a Data Conversion transformation step in between the Excel source and OLE DB destination. The meta data of the Excel source output has set the data type of policy_ref to DT_R8 (a float type) and my Data Conversion creates a copy of this column to a DT_STR (varchar), which matches my OLE DB destination. After the package has executed I find that all but 53 records on my OLE DB destination table have a policy_ref value of NULL despite the facts that no blanks or NULLs exist in the Excel source for policy_ref.
Can anyone advise why this behaviour is occurring and how to manage it properly within SSIS? I realise that I could explicitly format the column in Excel, which may do the trick, but this is not really the solution I'm looking for because I want to automate the SSIS package and need to cater for users other than me updating the Excel source in ways that might remove a text format from the policy_ref column. If formatting the Excel spreadsheet first is the only way then so be it but I'd appreciate expert input on this before reaching such a conclusion.
Many thanks in advance.