Faster hash maps in R
Contains a fix for a bug in 0.2.1 which caused a build error on macOS.
spp::sparse_hash_map
(from the C++ library sparsepp
), rather than boost::unordered_map
. In all cases this should result in improved performance and reduced memory usage; for a more detailed analysis refer to the sparsepp benchmark. To revert to boost::unordered_map
on these platforms, add -DHASHMAP_NO_SPP
to PKG_CPPFLAGS
in the Makevars file and rebuild the package.Added save_hashmap
and load_hashmap
function for saving and loading
Hashmap
objects (#3).
Added S3 plot
method.
Added clone
function for creating deep copies. This is also reflected in
the C++ API via HashMap::clone
.
Added S3 merge
method
Integers are now printed correctly.
Key and value vector access is now automatically cached on the first call to
$keys()
and $values()
, respectively.
Better handling of Date
and POSIXt
names for $data()
(#4).