Creates or validates a subclass of a tibble. These function is mostly useful for package authors that implement subclasses of a tibble, like sf or tsibble.

new_tibble() creates a new object as a subclass of tbl_df, tbl and data.frame. This function is optimized for performance, checks are reduced to a minimum.

validate_tibble() checks a tibble for internal consistency. Correct behavior can be guaranteed only if this function runs without raising an error.

new_tibble(x, ..., nrow, class = NULL, subclass = NULL)




A tibble-like object.


Name-value pairs of additional attributes.


The number of rows, required.


Subclasses to assign to the new object, default: none.


Deprecated, retained for compatibility. Please use the class argument.


For new_tibble(), x must be a list. The ... argument allows adding more attributes to the subclass. An nrow argument is required. This should be an integer of length 1, and every element of the list x should have vctrs::vec_size() equal to this value. (But this is not checked by the constructor). This takes the place of the "row.names" attribute in a data frame. x must have names (or be empty), but the names are not checked for correctness.


validate_tibble() checks for "minimal" names and that all columns are vectors, data frames or matrices. It also makes sure that all columns have the same length, and that vctrs::vec_size() is consistent with the data.

See also

tibble() and as_tibble() for ways to construct a tibble with recycling of scalars and automatic name repair.


# The nrow argument is always required:
new_tibble(list(a = 1:3, b = 4:6), nrow = 3)
#> # A tibble: 3 x 2
#>       a     b
#>   <int> <int>
#> 1     1     4
#> 2     2     5
#> 3     3     6

# Existing row.names attributes are ignored:
try(validate_tibble(new_tibble(iris, nrow = 3)))
#> Error : Tibble columns must have compatible sizes.
#> * Size 3: Requested with `nrow` argument.
#> * Size 150: Columns `Sepal.Length`, `Sepal.Width`, `Petal.Length`, `Petal.Width`, and `Species`.
#>  Only values of size one are recycled.

# The length of all columns must be compatible with the nrow argument:
try(validate_tibble(new_tibble(list(a = 1:3, b = 4:6), nrow = 2)))
#> Error : Tibble columns must have compatible sizes.
#> * Size 2: Requested with `nrow` argument.
#> * Size 3: Columns `a` and `b`.
#>  Only values of size one are recycled.