Reshaping data using tidyr::pivot_wider — pivot_wider.trackr

A drop in replacement for tidyr::pivot_wider() which optionally takes a message and headline to store in the history graph.

Usage

# S3 method for class 'trackr_df'
pivot_wider(
  data,
  ...,
  id_cols = NULL,
  id_expand = FALSE,
  names_from = name,
  names_prefix = "",
  names_sep = "_",
  names_glue = NULL,
  names_sort = FALSE,
  names_vary = "fastest",
  names_expand = FALSE,
  names_repair = "check_unique",
  values_from = value,
  values_fill = NULL,
  values_fn = NULL,
  unused_fn = NULL,
  .messages = c("wide format", "{.count.in} before", "{.count.out} after"),
  .headline = .defaultHeadline()
)

Arguments

data

A data frame to pivot.

...

Additional arguments passed on to methods.

id_cols

<tidy-select> A set of columns that uniquely identify each observation. Typically used when you have redundant variables, i.e. variables whose values are perfectly correlated with existing variables.

Defaults to all columns in data except for the columns specified through names_from and values_from. If a tidyselect expression is supplied, it will be evaluated on data after removing the columns specified through names_from and values_from.

id_expand

Should the values in the id_cols columns be expanded by expand() before pivoting? This results in more rows, the output will contain a complete expansion of all possible values in id_cols. Implicit factor levels that aren't represented in the data will become explicit. Additionally, the row values corresponding to the expanded id_cols will be sorted.

names_from, values_from

<tidy-select> A pair of arguments describing which column (or columns) to get the name of the output column (names_from), and which column (or columns) to get the cell values from (values_from).

If values_from contains multiple values, the value will be added to the front of the output column.

names_prefix

String added to the start of every variable name. This is particularly useful if names_from is a numeric vector and you want to create syntactic variable names.

names_sep

If names_from or values_from contains multiple variables, this will be used to join their values together into a single string to use as a column name.

names_glue

Instead of names_sep and names_prefix, you can supply a glue specification that uses the names_from columns (and special .value) to create custom column names.

names_sort

Should the column names be sorted? If FALSE, the default, column names are ordered by first appearance.

names_vary

When names_from identifies a column (or columns) with multiple unique values, and multiple values_from columns are provided, in what order should the resulting column names be combined?

"fastest" varies names_from values fastest, resulting in a column naming scheme of the form: value1_name1, value1_name2, value2_name1, value2_name2. This is the default.
"slowest" varies names_from values slowest, resulting in a column naming scheme of the form: value1_name1, value2_name1, value1_name2, value2_name2.

names_expand

Should the values in the names_from columns be expanded by expand() before pivoting? This results in more columns, the output will contain column names corresponding to a complete expansion of all possible values in names_from. Implicit factor levels that aren't represented in the data will become explicit. Additionally, the column names will be sorted, identical to what names_sort would produce.

names_repair

What happens if the output has invalid column names? The default, "check_unique" is to error if the columns are duplicated. Use "minimal" to allow duplicates in the output, or "unique" to de-duplicated by adding numeric suffixes. See vctrs::vec_as_names() for more options.

values_fill

Optionally, a (scalar) value that specifies what each value should be filled in with when missing.

This can be a named list if you want to apply different fill values to different value columns.

values_fn

Optionally, a function applied to the value in each cell in the output. You will typically use this when the combination of id_cols and names_from columns does not uniquely identify an observation.

This can be a named list if you want to apply different aggregations to different values_from columns.

unused_fn

Optionally, a function applied to summarize the values from the unused columns (i.e. columns not identified by id_cols, names_from, or values_from).

The default drops all unused columns from the result.

This can be a named list if you want to apply different aggregations to different unused columns.

id_cols must be supplied for unused_fn to be useful, since otherwise all unspecified columns will be considered id_cols.

This is similar to grouping by the id_cols then summarizing the unused columns using unused_fn.

.messages

a set of glue specs. The glue code can use any global variable, grouping variable, or {.strata}. Defaults to nothing.

.headline

a headline glue spec. The glue code can use any global variable, grouping variable, or {.strata}. Defaults to nothing.

Value

the data dataframe result of the tidyr::pivot_wider function but with a history graph updated.

Examples

library(dplyr)
library(dtrackr)

starwars %>%
  track() %>%
  tidyr::unnest(starships, keep_empty = TRUE) %>%
  tidyr::nest(world_data = c(-homeworld)) %>%
  history()
#> dtrackr history:
#> number of flowchart steps: 2 (approx)
#> tags defined: <none>
#> items excluded so far: <not capturing exclusions>
#> last entry / entries:
#> └ "13 items"

# There is a problem with `tidyr::unnest` that means if you want to override the
# `.messages` option at the moment it will most likely fail. Forcing the use of
# the specific `dtrackr::p_unnest` version solves this problem, until hopefully it is
# resolved in `tidyr`:
starwars %>%
  track() %>%
  p_unnest(
    films,
    .messages = c("{.count.in} characters", "{.count.out} appearances")
  ) %>%
  dplyr::group_by(gender) %>%
  tidyr::nest(
    people = c(-gender, -species, -homeworld),
    .messages = c("{.count.in} appearances", "{.count.out} planets")
  ) %>%
  status() %>%
  history()
#> dtrackr history:
#> number of flowchart steps: 5 (approx)
#> tags defined: <none>
#> items excluded so far: <not capturing exclusions>
#> last entry / entries:
#> ├ [NA]: "3 items"
#> ├ [feminine]: "feminine", "13 items"
#> └ [masculine]: "masculine", "49 items"

# This example includes pivoting and nesting. The CMS patient care data
# has multiple tests per institution in a long format, and observed /
# denominator types. Firstly we pivot the data to allow us to easily calculate
# a total percentage for each institution. This is duplicated for every test
# so we nest the tests to get to one row per institution. Those institutions
# with invalid scores are excluded.
cms_history = tidyr::cms_patient_care %>%
  track() %>%
  tidyr::pivot_wider(names_from = type, values_from = score) %>%
  dplyr::mutate(
    percentage = sum(observed) / sum(denominator) * 100,
    .by = c(ccn, facility_name)
  ) %>%
  tidyr::nest(
    results = c(measure_abbr, observed, denominator),
    .messages = c("{.count.in} test results", "{.count.out} facilities")
  ) %>%
  exclude_all(
    percentage > 100 ~ "{.excluded} facilities with anomalous percentages",
    na.rm = TRUE
  )

print(cms_history %>% dtrackr::history())
#> dtrackr history:
#> number of flowchart steps: 3 (approx)
#> tags defined: <none>
#> items excluded so far: <not capturing exclusions>
#> last entry / entries:
#> └ "126 test results", "14 facilities"

# not run in examples:
if (interactive()) {
  cms_history %>% flowchart()
}

Reshaping data using `tidyr::pivot_wider`

Usage

Arguments

Value

See also

Examples