Thanks to all for the suggestions. Pierre's approach (using Excel to read from datastream then R to read from Excel) seems like the most practical way, assuming one stays within the Excel spreadsheet size limits (I'm mainly interested in economic data series, so this shouldn't be a constraint).