[moderator: my earlier email was being held b/c of a mistake I made on
my end with my email address, sorry for the re-post]
I personally use a mysql database here with the 'quantmod' package. See:
?getSymbols.MySQL for the nitty gritty details.
You could also just open a connection to mysql either directly or via
ODBC and convert your time series into a data.frame and then use
dbWriteTable(dbcon, "SymbolName", Symbol.df). Conversely, to read back
into R, just use dbReadTable(dbcon, "SymbolName") and convert it into a
xts time series.
?dbReadTable and ?dbWriteTable ... If your dataset only adds the new row
everyday, when writing the table you could specify append=TRUE,
overwrite=FALSE.
HTH,
C