Support for reading metadata when writing a DataTable to CSV?

Dec 9, 2011 at 3:19 AM
Edited Dec 9, 2011 at 3:21 AM

Is it possible at present to have delimiting of values dependent on their type? So for example, string values are always delimited, but numeric or boolean values are not? At the moment, you can either have smart delimiting (the value starts/ends with a space, or contains the value separator or delimiter character or CR or LF characters) or delimit everything. If we're exporting from a DataTable then we have a lot of metadata available that could be checked.

The problem appears to be that the DataRecord converts all values read from a row into a collection of strings, so all the metadata is discarded.