Introducing <percent> vectors, a re-imagined and
simplified approach to working with percentages in R.
It is implemented as a lightweight S3 object that automatically formats proportions as percentages, simplifying any workflow that uses percentages at any stage of analysis.
Rather than:
x <- c(0.25, 0.5, 0.75)
paste0(x * 100, "%")## [1] "25%" "50%" "75%"You can now instead write:
x <- c(0.25, 0.5, 0.75)
as_percent(x)## [1] "25%" "50%" "75%"as_percent() keeps the proportions as-is and simply
prints them as percentages.
p <- as_percent(0.000567)
p## [1] "0.06%"as.double(p) # Under-the-hood nothing is modified## [1] 0.000567This allows us to use <percent> vectors throughout
our script without ever needing to convert to a character vector of
percentages.
Maths operations are also possible with <percent>
such as multiplication, division, rounding, summary statistics, etc. The
more traditional workflow would require a lot more effort to do these
same operations.
A percent vignette has been created to help introduce and explain this new feature.
file_size() has been deprecated as we believe it is no
longer relevant (#11).
Please let us know if you still have a use for it, otherwise it will be
removed in a future version.age_calculate() is now more robust in uncommon
situations such as leap years, or if a date is supplied as a date time.
Previously, it would sometimes give incorrect values for these types of
edge cases.postcode() is now
format_postcode(); age_group() is now
create_age_groups() and fin_year() is now
extract_fin_year().age_from_chi() due to a helper function
expected_age() caused the tests fail when the new year
comes. It has been replaced with a fixed reference date.Fix a bug in extract_fin_year() to make sure
financial years are displayed correctly from 1999/20 to
2008/09.
A new article has been added to the documentation - Working with CHI numbers.
The {gdata} import has been dropped and replaced with {scales}.
extract_fin_year() is now much faster and will use
less memory, especially for smaller vectors (1 to 1,000).
format_postcode() is now faster and also gains a
quiet parameter, the default value is FALSE
but setting it to TRUE will skip some of the checks and
messages, this is useful when using format_postcode() to
‘clean-up’ and format a vector of postcodes, rather than wanting to
check them. Because of the skipped checks quiet = TRUE
should also run faster.
The installation instructions in the README have been updated.
Improved chi_check() to make it more efficient and
run faster.
Improved the “Using phsmethods” section in the README to be shorter and more accessible.
Update all errors, warnings and messages to use {cli}.
Improve errors when giving incorrect types to some functions.
Three functions renamed to improve code clarity:
postcode() to format_postcode();
age_group() to create_age_groups();
fin_year() to extract_fin_year(). The old
functions will still work but will produce a warning. After a reasonable
amount of time, they will be removed completely.
New functions added: age_calculate()(#65,
@Nic-chr);
dob_from_chi()(#42,
@Moohan); and
age_from_chi()(#42,
@Moohan)
Change the output for chi_check so that empty string
(““) reports as missing (#76)
New functions added: age_group()(#23,
@chrisdeans);
chi_check()(#30,
@graemegowans);
chi_pad()(#30,
@graemegowans); and
match_area()(#13,
@jvillacampa).
The first argument of postcode() is now
x, as opposed to string. This is unlikely to
break much, if any, existing code. postcode() is also now
slightly faster.
phsmethods no longer imports
stringi.
phsmethods now depends on a version of R >=
2.10.
Jack Hannah is no longer a maintainer.
file_size(), fin_year(),
qtr(), qtr_end(), qtr_next() and
qtr_prev() now use inherits(x, "y") instead of
class(x) == "y" to check class. The reasoning is explained
in this blogpost
by Martin Maechler.
The performance of fin_year() has been improved. The
function now extracts the unique date(s) from the input, calculates the
associated financial year(s), and joins to the original input. This is
in contrast with the original method, which directly calculated the
financial year of all input dates individually.
file_size(), fin_year(),
postcode(), qtr(), qtr_end(),
qtr_next() and qtr_prev() functions
added.