keys-set {keyholder}

Set keys

Description

Key is a vector which goal is to provide information about rows in reference data frame. Its length should always be equal to number of rows in data frame. Keys are stored as tibble in attribute "keys" and so one data frame can have multiple keys. Data frame with keys is implemented as class keyed_df.

Usage

keys(.tbl) <- value

assign_keys(.tbl, value)

key_by(.tbl, ..., .add = FALSE, .exclude = FALSE)

unkey(.tbl)

Arguments

.tbl

Reference data frame .

value

Values of keys (converted to tibble).

...

Variables to be used as keys defined in similar fashion as in dplyr::select().

.add

Whether to add keys to (possibly) existing ones. If FALSE keys will be overridden.

.exclude

Whether to exclude key variables from .tbl.

Details

key_by ignores grouping when creating keys. Also if .add == TRUE and names of some added keys match the names of existing keys the new ones will override the old ones.

Value for ⁠keys<-⁠ should not be NULL because it is converted to tibble with zero rows. To remove keys use unkey(), remove_keys() or restore_keys(). assign_keys is a more suitable for piping wrapper for ⁠keys<-⁠.

See Also

Get keys, Manipulate keys

Scoped key_by()

Examples

df <- dplyr::as_tibble(mtcars)

# Value is converted to tibble
keys(df) <- 1:nrow(df)

# This will throw an error
## Not run: 
keys(df) <- 1:10

## End(Not run)

# Use 'vs' and 'am' as keys
df %>% key_by(vs, am)

df %>% key_by(vs, am, .exclude = TRUE)

df %>% key_by(vs) %>% key_by(am, .add = TRUE, .exclude = TRUE)

# Override keys
df %>% key_by(vs, am) %>% dplyr::mutate(vs = 1) %>%
  key_by(gear, vs, .add = TRUE)

# Use select helpers
df %>% key_by(dplyr::one_of(c("vs", "am")))

df %>% key_by(dplyr::everything())