Translation of missing numbers between DSS and R

[eheisman]

DSS uses a special value (~ -3.4e38) to represent missing values. Right now these are added back into the data where timestamps are missing when exporting to a DSS time series.

On import, nothing special is done to remove these values. Sometimes they will not occur in a time series if the entire record is missing.

Solution: Add functionality in xts.to.tsc and tsc.to.xts to convert between the DSS undefined value constant and NAs in the time series, or drop values with those timestamps on import.