NEWS.md
stype 0.5.0outcomes (or censors) and all the vectors have
short labels, then the list names are the levels and the short labels
become the labels.outcomes (or
censors) and any of the vectors are missing short
labels, then the list names become the levels and the labels.as.character(1:length(x)) become the levels
and labels, where length(x) is the number of list
elements.v_continuous and
v_continuous_nonneg as the w argument to
weight possible.v_rcensored to catch invalid
inputs and provide more helpful error messages.stype under which a
stype vector was created as an attribute. This is included
for future compatibility checking when comparing vectors in the case
they were created under different versions of stype that
don’t play nice together.new_stype_vctr function to streamline
the creation of new stype vectors, ensuring they all share a common set
of attributes.v_character type.v_proportion types with values in
[0. 1], which inherits from
v_continuous_nonneg.v_event_time type.v_rcensored type now takes
v_continuous_nonneg vectors instead of
v_event_time. More importantly, NA values are
no longer accepted as inputs to v_rcensored. To indicate
that an observation has not yet been observed use
Inf.auto_compute_summary argument to stype
v_* constructors. When TRUE, data summaries
are automatically computed when a vector is created or restored (e.g. on
subset or c()). When FALSE, data summaries are
only computed when the user asks for a data summary. The default is
TRUE, though this may change in the future.v_continous, v_continuous_nonneg,
v_binary, v_proportion, and
v_count. See the stype math vignette for a
complete list of all available functions and the respective
domains/codomains of the functions.ptableNoNA descriptor to
categoricalDescriptors which summarizes a table excluding
NA values.stype 0.4.3as_Surv to take a
new formal argument multiple_endpoint which defaults to
FALSE. Previous versions always returned data in a
multiple endpoints format (see the object documentation for
survival::Surv for a definition) when
censor_as_event was FALSE, so this is a
breaking change.pmin_v_rcensored that can be used to
combine multiple v_rcensored objects into a new
v_rcensored object.stype 0.4.2bnry : v_binary
cnt : v_count
chr : v_character
cont : v_continuous
nneg : v_continuous_nonneg
tmev : v_event_time
nom : v_nominal
ord : v_ordered
rcen : v_rcensored
stype 0.4.1* and + where (e.g)
v_binary(c(FALSE, TRUE, FALSE, TRUE)) + c(1, 3, 3, 5) now
returns an error.stype 0.4.0tbl_analysis type, which
includes the basic utilities of a stype such as a
context and data_summary. It also includes a
modifiers argument which is a list of
functions sequentially modify the table. These functions are applied
whenever the table is subset, and thus can act to define analytic
workflows that should be run whenever a table is modified. This feature
is experimental and will be tested internally at NoviSci before being
further developed.stype 0.3.0vctrs 0.3.0.stypes have an
extra_descriptors argument for passing a
listof descriptor function.purpose class now has a tags slot, so
stype vectors can be tagged with arbitrary
character vectors. The is_tagged function can
be used to check if a vector is tagged with a string;
e.g. is_tagged(x, tags = c("tag1", "tag2")) checks if
x has either "tag1" or
"tag2".lenses! See the lenses vignette for examples of getting/setting
parts of stype vector or getting/setting several
stypes within a list-like structure.weight function is available for updating the
data_summary of stypes with vector of weights.
See the weighting vignette for example of
usage.get_data_summaries function to access the
data_summary from multiple stypes in a list or
data.frame.