Map data records from block records (where each record may be more than one row) to row records (where each record is a single row). Values specified in rowKeyColumns determine which sets of rows build up records and are copied into the result.

pivot_to_rowrecs(
  data,
  columnToTakeKeysFrom,
  columnToTakeValuesFrom,
  rowKeyColumns,
  ...,
  sep = NULL,
  checkNames = TRUE,
  checkKeys = TRUE,
  strict = FALSE,
  allow_rqdatatable = FALSE
)

layout_to_rowrecs(
  data,
  columnToTakeKeysFrom,
  columnToTakeValuesFrom,
  rowKeyColumns,
  ...,
  sep = NULL,
  checkNames = TRUE,
  checkKeys = TRUE,
  strict = FALSE,
  allow_rqdatatable = FALSE
)

Arguments

data

data.frame to work with (must be local, for remote please try moveValuesToColumns*).

columnToTakeKeysFrom

character name of column build new column names from.

columnToTakeValuesFrom

character name of column to get values from.

rowKeyColumns

character array names columns that should be table keys.

...

force later arguments to bind by name.

sep

character if not null build more detailed column names.

checkNames

logical, if TRUE check names.

checkKeys

logical, if TRUE check keyColumns uniquely identify blocks (required).

strict

logical, if TRUE check control table name forms

allow_rqdatatable

logical, if TRUE allow rqdatatable shortcutting on simple conversions.

Value

new data.frame with values moved to columns.

See also

Examples

d <- data.frame(model_id = c("m1", "m1"), meas = c('AUC', 'R2'), val= c(0.6, 0.2)) pivot_to_rowrecs(d, columnToTakeKeysFrom= 'meas', columnToTakeValuesFrom= 'val', rowKeyColumns= "model_id") %.>% print(.)
#> model_id AUC R2 #> 1 m1 0.6 0.2