The IfxMRData data type
The main advantage of using the multirepresentational data type IfxMRData is that it dynamically determines whether to store your document in an LVARCHAR data type or a CLOB data type, depending on the size of the document.
This feature is useful if the documents you want to store in a single column of a table vary greatly in size, where some fit in an LVARCHAR data type and some do not. It is preferable to store documents in LVARCHAR columns if at all possible, since it is faster to insert documents into and retrieve documents from this data type than into smart large object types such as CLOB. IfxMRData lets the data type itself decide where to store the data so that you do not have to decide at the time you create the table.
IfxMRData is also designed to improve I/O performance of documents that contain only standard ASCII text. Therefore, if your documents do not contain binary data, storing them in a column of this data type might improve performance.