qs2: a framework for efficient serialization
qs2
will be the successor the qs
package. The goal is to have
cutting edge serialization performance in R.
Two new formats are introduced: the qs2
and qdata
formats.
The qs2
format directly uses R’s serialization functions, so it should
be fully compatible with all future version of R. For most people, this
is what you want to use. It’s a drop in replacement for qs::qsave
or
saveRDS
.
The qdata
format has its own layout and works with only R data types
(vectors, lists, data frames, matrices). It will replace internal types
(functions, promises, external pointers, environments) with NULL. This
has slightly better performance but is not general.
library(qs2)
qs_save(data, "myfile.qs2")
data <- qs_read("myfile.qs2")
qd_save(data, "myfile.qd")
data <- qd_read("myfile.qd")
TO DO: description of datasets used.
See inst/benchmarks
for more details and a comparison to other
libraries.