The data storage is in ascii format and so each retrieval causes a reparsing.
It is much better to store in native format so that this parsing overhead is
avoided altogether. The data would be stored in a binary blob type
The meta data associated with such regular time series is further not
associated via fields such as units, sampling type. These should be added as
a field to the type.
Original issue reported on code.google.com by sandhuda...@gmail.com on 30 Dec 2009 at 12:52
Original issue reported on code.google.com by
sandhuda...@gmail.com
on 30 Dec 2009 at 12:52