Skip to contents

Check table relations

Source: R/check-cardinalities.R
examine_cardinality.Rd

All check_cardinality_...() functions test the following conditions:

  1. Are all rows in x unique?

  2. Are the rows in y a subset of the rows in x?

  3. Does the relation between x and y meet the cardinality requirements? One row from x must correspond to the requested number of rows in y, e.g. _0_1 means that there must be zero or one rows in y for each row in x.

examine_cardinality() also checks the first two points and subsequently determines the type of cardinality.

For convenience, the x_select and y_select arguments allow restricting the check to a set of key columns without affecting the return value.

Usage

check_cardinality_0_n(
  x,
  y,
  ...,
  x_select = NULL,
  y_select = NULL,
  by_position = NULL
)

check_cardinality_1_n(
  x,
  y,
  ...,
  x_select = NULL,
  y_select = NULL,
  by_position = NULL
)

check_cardinality_1_1(
  x,
  y,
  ...,
  x_select = NULL,
  y_select = NULL,
  by_position = NULL
)

check_cardinality_0_1(
  x,
  y,
  ...,
  x_select = NULL,
  y_select = NULL,
  by_position = NULL
)

examine_cardinality(
  x,
  y,
  ...,
  x_select = NULL,
  y_select = NULL,
  by_position = NULL
)

Arguments

x

Parent table, data frame or lazy table.

y

Child table, data frame or lazy table.

...

These dots are for future extensions and must be empty.

x_select, y_select

Key columns to restrict the check, processed with dplyr::select().

by_position

Set to TRUE to ignore column names and match by position instead. The default means matching by name, use x_select and/or y_select to align the names.

Value

check_cardinality_...() return x, invisibly, if the check is passed, to support pipes. Otherwise an error is thrown and the reason for it is explained.

examine_cardinality() returns a character variable specifying the type of relationship between the two columns.

Details

All cardinality functions accept a parent and a child table (x and y). All rows in x must be unique, and all rows in y must be a subset of the rows in x. The x_select and y_select arguments allow restricting the check to a set of key columns without affecting the return value. If given, both arguments must refer to the same number of key columns.

The cardinality specifications "0_n", "1_n", "0_1", "1_1" refer to the expected relation that the child table has with the parent table. "0", "1" and "n" refer to the occurrences of value combinations in y that correspond to each combination in the columns of the parent table. "n" means "more than one" in this context, with no upper limit.

"0_n": no restrictions, each row in x has at least 0 and at most n corresponding occurrences in y.

"1_n": each row in x has at least 1 and at most n corresponding occurrences in y. This means that there is a "surjective" mapping from the child table to the parent table, i.e. each parent table row exists at least once in the child table.

"0_1": each row in x has at least 0 and at most 1 corresponding occurrence in y. This means that there is a "injective" mapping from the child table to the parent table, i.e. no combination of values in the parent table columns is addressed multiple times. But not all parent table rows have to be referred to.

"1_1": each row in x occurs exactly once in y. This means that there is a "bijective" ("injective" AND "surjective") mapping between the child table and the parent table, i.e. the sets of rows are identical.

Finally, examine_cardinality() tests for and returns the nature of the relationship (injective, surjective, bijective, or none of these) between the two given sets of columns. If either x is not unique or there are rows in y that are missing from x, the requirements for a cardinality test is not fulfilled. No error will be thrown, but the result will contain the information which prerequisite was violated.

See also

Other cardinality functions: dm_examine_cardinalities()

Examples

d1 <- tibble::tibble(a = 1:5)
d2 <- tibble::tibble(a = c(1:4, 4L))
d3 <- tibble::tibble(c = c(1:5, 5L), d = 0)
# This does not pass, `a` is not unique key of d2:
try(check_cardinality_0_n(d2, d1))
#> Error in abort_not_unique_key(x_label, orig_names) : 
#>   (`a`) not a unique key of `d2`.

# Columns are matched by name by default:
try(check_cardinality_0_n(d1, d3))
#> Error in check_card_api_impl({ : 
#>   `by_position = FALSE` or `by_position = NULL` require column names in `x` to match those in `y`.

# This passes, multiple values in d3$c are allowed:
check_cardinality_0_n(d1, d2)

# This does not pass, injectivity is violated:
try(check_cardinality_1_1(d1, d3, y_select = c(a = c)))
#> Error in abort_not_bijective(y_label, colnames(y)) : 
#>   1..1 cardinality (bijectivity) is not given: Column (`a`) in table `d3` contains duplicate values.
try(check_cardinality_0_1(d1, d3, x_select = c(c = a)))
#> Error in abort_not_injective(y_label, colnames(y)) : 
#>   0..1 cardinality (injectivity from child table to parent table) is not given: Column (`c`) in table `d3` contains duplicate values.

# What kind of cardinality is it?
examine_cardinality(d1, d3, x_select = c(c = a))
#> [1] "surjective mapping (child: 1 to n -> parent: 1)"
examine_cardinality(d1, d2)
#> [1] "generic mapping (child: 0 to n -> parent: 1)"