ivs: Interval Vectors

Description

Provides a library for generic interval manipulations using a new interval vector class. Capabilities include: locating various kinds of relationships between two interval vectors, merging overlaps within a single interval vector, splitting an interval vector on its overlapping endpoints, and applying set theoretical operations on interval vectors. Many of the operations in this package were inspired by James Allen's interval algebra, Allen (1983) doi:10.1145/182.358434.

Author(s)

Maintainer: Davis Vaughan davis@posit.co

Other contributors:

• Posit Software, PBC [copyright holder, funder]

See Also

Useful links:

• https://github.com/DavisVaughan/ivs

• https://davisvaughan.github.io/ivs/

• Report bugs at https://github.com/DavisVaughan/ivs/issues

Count relations from Allen's Interval Algebra

Description

iv_count_relates() is similar to iv_count_overlaps(), but it counts a specific set of relations developed by James Allen in the paper: Maintaining Knowledge about Temporal Intervals.

Usage

iv_count_relates(
  needles,
  haystack,
  ...,
  type,
  missing = "equals",
  no_match = 0L
)

Arguments

Value

An integer vector the same size as needles.

Allen's Interval Algebra

The interval algebra developed by James Allen serves as a basis and inspiration for iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows(). The original algebra is composed of 13 relations which have the following properties:

• Distinct: No pair of intervals can be related by more than one type.

• Exhaustive: All pairs of intervals are described by one of the types.

• Qualitative: No numeric intervals are considered. The relationships are computed by purely qualitative means.

Take the notation that x and y represent two intervals. Now assume that x can be represented as ⁠[x_s, x_e)⁠, where x_s is the start of the interval and x_e is the end of it. Additionally, assume that x_s < x_e. With this notation, the 13 relations are as follows:

• Precedes:

  x_e < y_s

• Preceded-by:

  x_s > y_e

• Meets:

  x_e == y_s

• Met-by:

  x_s == y_e

• Overlaps:

  (x_s < y_s) & (x_e > y_s) & (x_e < y_e)

• Overlapped-by:

  (x_e > y_e) & (x_s < y_e) & (x_s > y_s)

• Starts:

  (x_s == y_s) & (x_e < y_e)

• Started-by:

  (x_s == y_s) & (x_e > y_e)

• Finishes:

  (x_s > y_s) & (x_e == y_e)

• Finished-by:

  (x_s < y_s) & (x_e == y_e)

• During:

  (x_s > y_s) & (x_e < y_e)

• Contains:

  (x_s < y_s) & (x_e > y_e)

• Equals:

  (x_s == y_s) & (x_e == y_e)

Note that when missing = "equals", missing intervals will only match the type = "equals" relation. This ensures that the distinct property of the algebra is maintained.

Connection to other functions

Note that some of the above relations are fairly restrictive. For example, "overlaps" only detects cases where x straddles y_s. It does not consider the case where x and y are equal to be an overlap (as this is "equals") nor does it consider when x straddles y_e to be an overlap (as this is "overlapped-by"). This makes the relations extremely useful from a theoretical perspective, because they can be combined without fear of duplicating relations, but they don't match our typical expectations for what an "overlap" is.

iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows() use more intuitive types that aren't distinct, but typically match your expectations better. They can each be expressed in terms of Allen's relations:

• iv_locate_overlaps():

  • "any":

    overlaps | overlapped-by | starts | started-by | finishes | finished-by | during | contains | equals

  • "contains":

    contains | started-by | finished-by | equals

  • "within":

    during | starts | finishes | equals

  • "starts":

    starts | started-by | equals

  • "ends":

    finishes | finished-by | equals

  • "equals":

    equals

• iv_locate_precedes():

  precedes | meets

• iv_locate_follows():

  preceded-by | met-by

See Also

Locating relations from Allen's Interval Algebra

Examples

x <- iv(1, 3)
y <- iv(3, 4)

# `"precedes"` is strict, and doesn't let the endpoints match
iv_count_relates(x, y, type = "precedes")

# Since that is what `"meets"` represents
iv_count_relates(x, y, type = "meets")

# `"overlaps"` is a very specific type of overlap where an interval in
# `needles` straddles the start of an interval in `haystack`
x <- iv_pairs(c(1, 4), c(1, 3), c(0, 3), c(2, 5))
y <- iv(1, 4)

# It doesn't match equality, or when the starts match, or when the end
# of the interval in `haystack` is straddled instead
iv_count_relates(x, y, type = "overlaps")

Detect relations from Allen's Interval Algebra

Description

iv_relates() is similar to iv_overlaps(), but it detects a specific set of relations developed by James Allen in the paper: Maintaining Knowledge about Temporal Intervals.

Usage

iv_relates(needles, haystack, ..., type, missing = "equals")

Arguments

Value

A logical vector the same size as needles.

Allen's Interval Algebra

The interval algebra developed by James Allen serves as a basis and inspiration for iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows(). The original algebra is composed of 13 relations which have the following properties:

• Distinct: No pair of intervals can be related by more than one type.

• Exhaustive: All pairs of intervals are described by one of the types.

• Qualitative: No numeric intervals are considered. The relationships are computed by purely qualitative means.

Take the notation that x and y represent two intervals. Now assume that x can be represented as ⁠[x_s, x_e)⁠, where x_s is the start of the interval and x_e is the end of it. Additionally, assume that x_s < x_e. With this notation, the 13 relations are as follows:

• Precedes:

  x_e < y_s

• Preceded-by:

  x_s > y_e

• Meets:

  x_e == y_s

• Met-by:

  x_s == y_e

• Overlaps:

  (x_s < y_s) & (x_e > y_s) & (x_e < y_e)

• Overlapped-by:

  (x_e > y_e) & (x_s < y_e) & (x_s > y_s)

• Starts:

  (x_s == y_s) & (x_e < y_e)

• Started-by:

  (x_s == y_s) & (x_e > y_e)

• Finishes:

  (x_s > y_s) & (x_e == y_e)

• Finished-by:

  (x_s < y_s) & (x_e == y_e)

• During:

  (x_s > y_s) & (x_e < y_e)

• Contains:

  (x_s < y_s) & (x_e > y_e)

• Equals:

  (x_s == y_s) & (x_e == y_e)

Note that when missing = "equals", missing intervals will only match the type = "equals" relation. This ensures that the distinct property of the algebra is maintained.

Connection to other functions

Note that some of the above relations are fairly restrictive. For example, "overlaps" only detects cases where x straddles y_s. It does not consider the case where x and y are equal to be an overlap (as this is "equals") nor does it consider when x straddles y_e to be an overlap (as this is "overlapped-by"). This makes the relations extremely useful from a theoretical perspective, because they can be combined without fear of duplicating relations, but they don't match our typical expectations for what an "overlap" is.

iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows() use more intuitive types that aren't distinct, but typically match your expectations better. They can each be expressed in terms of Allen's relations:

• iv_locate_overlaps():

  • "any":

    overlaps | overlapped-by | starts | started-by | finishes | finished-by | during | contains | equals

  • "contains":

    contains | started-by | finished-by | equals

  • "within":

    during | starts | finishes | equals

  • "starts":

    starts | started-by | equals

  • "ends":

    finishes | finished-by | equals

  • "equals":

    equals

• iv_locate_precedes():

  precedes | meets

• iv_locate_follows():

  preceded-by | met-by

See Also

Locating relationships

Locating relations from Allen's Interval Algebra

Detecting relations from Allen's Interval Algebra pairwise

Examples

x <- iv(1, 3)
y <- iv(3, 4)

# `"precedes"` is strict, and doesn't let the endpoints match
iv_relates(x, y, type = "precedes")

# Since that is what `"meets"` represents
iv_relates(x, y, type = "meets")

# `"overlaps"` is a very specific type of overlap where an interval in
# `needles` straddles the start of an interval in `haystack`
x <- iv_pairs(c(1, 4), c(1, 3), c(0, 3), c(2, 5))
y <- iv(1, 4)

# It doesn't match equality, or when the starts match, or when the end
# of the interval in `haystack` is straddled instead
iv_relates(x, y, type = "overlaps")

Pairwise detect relations from Allen's Interval Algebra

Description

iv_pairwise_relates() is similar to iv_pairwise_overlaps(), but it detects a specific set of relations developed by James Allen in the paper: Maintaining Knowledge about Temporal Intervals.

Usage

iv_pairwise_relates(x, y, ..., type)

Arguments

Value

A logical vector the same size as the common size of x and y.

Allen's Interval Algebra

The interval algebra developed by James Allen serves as a basis and inspiration for iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows(). The original algebra is composed of 13 relations which have the following properties:

• Distinct: No pair of intervals can be related by more than one type.

• Exhaustive: All pairs of intervals are described by one of the types.

• Qualitative: No numeric intervals are considered. The relationships are computed by purely qualitative means.

Take the notation that x and y represent two intervals. Now assume that x can be represented as ⁠[x_s, x_e)⁠, where x_s is the start of the interval and x_e is the end of it. Additionally, assume that x_s < x_e. With this notation, the 13 relations are as follows:

• Precedes:

  x_e < y_s

• Preceded-by:

  x_s > y_e

• Meets:

  x_e == y_s

• Met-by:

  x_s == y_e

• Overlaps:

  (x_s < y_s) & (x_e > y_s) & (x_e < y_e)

• Overlapped-by:

  (x_e > y_e) & (x_s < y_e) & (x_s > y_s)

• Starts:

  (x_s == y_s) & (x_e < y_e)

• Started-by:

  (x_s == y_s) & (x_e > y_e)

• Finishes:

  (x_s > y_s) & (x_e == y_e)

• Finished-by:

  (x_s < y_s) & (x_e == y_e)

• During:

  (x_s > y_s) & (x_e < y_e)

• Contains:

  (x_s < y_s) & (x_e > y_e)

• Equals:

  (x_s == y_s) & (x_e == y_e)

Note that when missing = "equals", missing intervals will only match the type = "equals" relation. This ensures that the distinct property of the algebra is maintained.

Connection to other functions

Note that some of the above relations are fairly restrictive. For example, "overlaps" only detects cases where x straddles y_s. It does not consider the case where x and y are equal to be an overlap (as this is "equals") nor does it consider when x straddles y_e to be an overlap (as this is "overlapped-by"). This makes the relations extremely useful from a theoretical perspective, because they can be combined without fear of duplicating relations, but they don't match our typical expectations for what an "overlap" is.

iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows() use more intuitive types that aren't distinct, but typically match your expectations better. They can each be expressed in terms of Allen's relations:

• iv_locate_overlaps():

  • "any":

    overlaps | overlapped-by | starts | started-by | finishes | finished-by | during | contains | equals

  • "contains":

    contains | started-by | finished-by | equals

  • "within":

    during | starts | finishes | equals

  • "starts":

    starts | started-by | equals

  • "ends":

    finishes | finished-by | equals

  • "equals":

    equals

• iv_locate_precedes():

  precedes | meets

• iv_locate_follows():

  preceded-by | met-by

See Also

Locating relationships

Locating relations from Allen's Interval Algebra

Detecting relations from Allen's Interval Algebra

Examples

x <- iv_pairs(c(1, 3), c(3, 5))
y <- iv_pairs(c(3, 4), c(6, 7))

# `"precedes"` is strict, and doesn't let the endpoints match
iv_pairwise_relates(x, y, type = "precedes")

# Since that is what `"meets"` represents
iv_pairwise_relates(x, y, type = "meets")

# `"during"` only matches when `x` is completely contained in `y`, and
# doesn't allow any endpoints to match
x <- iv_pairs(c(1, 3), c(4, 5), c(8, 9))
y <- iv_pairs(c(1, 4), c(3, 8), c(8, 9))

iv_pairwise_relates(x, y, type = "during")

Locate relations from Allen's Interval Algebra

Description

iv_locate_relates() is similar to iv_locate_overlaps(), but it locates a specific set of relations developed by James Allen in the paper: Maintaining Knowledge about Temporal Intervals.

Usage

iv_locate_relates(
  needles,
  haystack,
  ...,
  type,
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

Arguments

Value

A data frame containing two integer columns named needles and haystack.

Allen's Interval Algebra

The interval algebra developed by James Allen serves as a basis and inspiration for iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows(). The original algebra is composed of 13 relations which have the following properties:

• Distinct: No pair of intervals can be related by more than one type.

• Exhaustive: All pairs of intervals are described by one of the types.

• Qualitative: No numeric intervals are considered. The relationships are computed by purely qualitative means.

Take the notation that x and y represent two intervals. Now assume that x can be represented as ⁠[x_s, x_e)⁠, where x_s is the start of the interval and x_e is the end of it. Additionally, assume that x_s < x_e. With this notation, the 13 relations are as follows:

• Precedes:

  x_e < y_s

• Preceded-by:

  x_s > y_e

• Meets:

  x_e == y_s

• Met-by:

  x_s == y_e

• Overlaps:

  (x_s < y_s) & (x_e > y_s) & (x_e < y_e)

• Overlapped-by:

  (x_e > y_e) & (x_s < y_e) & (x_s > y_s)

• Starts:

  (x_s == y_s) & (x_e < y_e)

• Started-by:

  (x_s == y_s) & (x_e > y_e)

• Finishes:

  (x_s > y_s) & (x_e == y_e)

• Finished-by:

  (x_s < y_s) & (x_e == y_e)

• During:

  (x_s > y_s) & (x_e < y_e)

• Contains:

  (x_s < y_s) & (x_e > y_e)

• Equals:

  (x_s == y_s) & (x_e == y_e)

Note that when missing = "equals", missing intervals will only match the type = "equals" relation. This ensures that the distinct property of the algebra is maintained.

Connection to other functions

Note that some of the above relations are fairly restrictive. For example, "overlaps" only detects cases where x straddles y_s. It does not consider the case where x and y are equal to be an overlap (as this is "equals") nor does it consider when x straddles y_e to be an overlap (as this is "overlapped-by"). This makes the relations extremely useful from a theoretical perspective, because they can be combined without fear of duplicating relations, but they don't match our typical expectations for what an "overlap" is.

iv_locate_overlaps(), iv_locate_precedes(), and iv_locate_follows() use more intuitive types that aren't distinct, but typically match your expectations better. They can each be expressed in terms of Allen's relations:

• iv_locate_overlaps():

  • "any":

    overlaps | overlapped-by | starts | started-by | finishes | finished-by | during | contains | equals

  • "contains":

    contains | started-by | finished-by | equals

  • "within":

    during | starts | finishes | equals

  • "starts":

    starts | started-by | equals

  • "ends":

    finishes | finished-by | equals

  • "equals":

    equals

• iv_locate_precedes():

  precedes | meets

• iv_locate_follows():

  preceded-by | met-by

References

Allen, James F. (26 November 1983). "Maintaining knowledge about temporal intervals". Communications of the ACM. 26 (11): 832–843.

See Also

Locating relationships

Detecting relations from Allen's Interval Algebra

Detecting relations from Allen's Interval Algebra pairwise

Examples

x <- iv(1, 3)
y <- iv(3, 4)

# `"precedes"` is strict, and doesn't let the endpoints match
iv_locate_relates(x, y, type = "precedes")

# Since that is what `"meets"` represents
iv_locate_relates(x, y, type = "meets")

# `"overlaps"` is a very specific type of overlap where an interval in
# `needles` straddles the start of an interval in `haystack`
x <- iv_pairs(c(1, 4), c(1, 3), c(0, 3), c(2, 5))
y <- iv(1, 4)

# It doesn't match equality, or when the starts match, or when the end
# of the interval in `haystack` is straddled instead
iv_locate_relates(x, y, type = "overlaps")

Is x an iv?

Description

is_iv() tests if x is an iv object.

Usage

is_iv(x)

Arguments

Value

A single TRUE or FALSE.

Examples

is_iv(1)
is_iv(new_iv(1, 2))

Create an interval vector

Description

• iv() creates an interval vector from start and end vectors. This is how you will typically create interval vectors, and is often used with columns in a data frame.

• iv_pairs() creates an interval vector from pairs. This is often useful for interactive testing, as it provides a more intuitive interface for creating small interval vectors. It should generally not be used on a large scale because it can be slow.

Intervals

Interval vectors are right-open, i.e. ⁠[start, end)⁠. This means that start < end is a requirement to generate an interval vector. In particular, empty intervals with start == end are not allowed.

Right-open intervals tend to be the most practically useful. For example, ⁠[2019-01-01 00:00:00, 2019-01-02 00:00:00)⁠ nicely encapsulates all times on 2019-01-01. With closed intervals, you'd have to attempt to specify this as ⁠2019-01-01 23:59:59⁠, which is inconvenient and inaccurate, as it doesn't capture fractional seconds.

Right-open intervals also have the extremely nice technical property that they create a closed algebra. Concretely, the complement of a vector of right-open intervals and the union, intersection, or difference of two vectors of right-open intervals will always result in another vector of right-open intervals.

Missing intervals

When creating interval vectors with iv(), if either bound is incomplete, then both bounds are set to their missing value.

Usage

iv(start, end, ..., ptype = NULL, size = NULL)

iv_pairs(..., ptype = NULL)

Arguments

Value

An iv.

Examples

library(dplyr, warn.conflicts = FALSE)

set.seed(123)

x <- tibble(
  start = as.Date("2019-01-01") + 1:5,
  end = start + sample(1:10, length(start), replace = TRUE)
)

# Typically you'll use `iv()` with columns of a data frame
mutate(x, iv = iv(start, end), .keep = "unused")

# `iv_pairs()` is useful for generating interval vectors interactively
iv_pairs(c(1, 5), c(2, 3), c(6, 10))

Access the start or end of an interval vector

Description

• iv_start() accesses the start of an interval vector.

• iv_end() accesses the end of an interval vector.

Usage

iv_start(x)

iv_end(x)

Arguments

Value

The start or end of x.

Examples

x <- new_iv(1, 2)

iv_start(x)
iv_end(x)

Containers

Description

This family of functions revolves around computing interval containers. A container is defined as the widest interval that isn't contained by any other interval.

• iv_containers() returns all of the containers found within x.

• iv_identify_containers() identifies the containers that each interval in x falls in. It replaces x with a list of the same size where each element of the list contains the containers that the corresponding interval in x falls in. This is particularly useful alongside tidyr::unnest().

• iv_identify_container() is similar in spirit to iv_identify_containers(), but is useful when you suspect that each interval in x is contained within exactly 1 container. It replaces x with an iv of the same size where each interval is the container that the corresponding interval in x falls in. If any interval falls in more than one container, an error is thrown.

• iv_locate_containers() returns a two column data frame with a key column containing the result of iv_containers() and a loc list-column containing integer vectors that map each interval in x to the container that it falls in.

Usage

iv_containers(x)

iv_identify_containers(x)

iv_identify_container(x)

iv_locate_containers(x)

Arguments

Value

• For iv_containers(), an iv with the same type as x.

• For iv_identify_containers(), a list-of containing ivs with the same size as x.

• For iv_identify_container(), an iv with the same type as x.

• For iv_locate_containers(), a two column data frame with a key column containing the result of iv_containers() and a loc list-column containing integer vectors.

Examples

library(dplyr, warn.conflicts = FALSE)
library(tidyr)

x <- iv_pairs(
  c(4, 6),
  c(1, 5),
  c(2, 3),
  c(NA, NA),
  c(NA, NA),
  c(9, 12),
  c(9, 14)
)
x

# Containers are intervals which aren't contained in any other interval.
# They are always returned in ascending order.
# If any missing intervals are present, a single one is retained.
iv_containers(x)

# `iv_identify_container()` is useful alongside `group_by()` and
# `summarize()` if you know that each interval is contained within exactly
# 1 container
df <- tibble(x = x)
df <- mutate(df, container = iv_identify_container(x))
df

df %>%
  group_by(container) %>%
  summarize(n = n())

# If any interval is contained within multiple containers,
# then you can't use `iv_identify_container()`
y <- c(x, iv_pairs(c(0, 3), c(8, 13)))
y

try(iv_identify_container(y))

# Instead, use `iv_identify_containers()` to identify every container
# that each interval falls in
df <- tibble(y = y, container = iv_identify_containers(y))
df

# You can use `tidyr::unchop()` to see the containers that each interval
# falls in
df %>%
  mutate(row = row_number(), .before = 1) %>%
  unchop(container)

# A more programmatic interface to `iv_identify_containers()` is
# `iv_locate_containers()`, which returns the containers you get from
# `iv_containers()` alongside the locations in the input that they contain.
iv_locate_containers(y)

Proxy and restore

Description

• iv_proxy() is an S3 generic which allows you to write S3 methods for iv extension types to ensure that they are treated like iv objects. The input will be your iv extension object, x, and the return value should be an iv object.

• iv_restore() is an S3 generic that dispatches off to that allows you to restore a proxied iv extension type back to its original type. The inputs will be a bare iv object, x, and your original iv extension object, to, and the return value should correspond to x restored to the type of to, if possible.

You typically don't need to create an iv_proxy() method if your class directly extends iv through the class argument of new_iv(). You only need to implement this if your class has a different structure than a typical iv object. In particular, if vctrs::field(x, "start") and vctrs::field(x, "end") don't return the start and end of the interval vector respectively, then you probably need an iv_proxy() method.

You typically do need an iv_restore() method for custom iv extensions. If your class is simple, then you can generally just call your constructor, like new_my_iv(), to restore the class and any additional attributes that might be required. If your class doesn't use new_iv(), then an iv_restore() method is mandatory, as this is one of the ways that ivs detects that your class is compatible with ivs.

This system allows you to use any ⁠iv_*()⁠ function on your iv extension object without having to define S3 methods for all of them.

Note that the default method for iv_proxy() returns its input unchanged, even if it isn't an iv. Each ⁠iv_*()⁠ function does separate checking to ensure that the proxy is a valid iv, or implements an alternate behavior if no proxy method is implemented. In contrast, iv_restore() will error if a method for to isn't registered.

Usage

iv_proxy(x, ...)

iv_restore(x, to, ...)

Arguments

Value

• iv_proxy() should return an iv object for further manipulation.

• iv_restore() should return an object of type to, if possible. In some cases, it may be required to fall back to returning an iv object.

Examples

if (FALSE) {
# Registering S3 methods outside of a package doesn't always work quite
# right (like on the pkgdown site), so this code should only be run by a
# user reading the manual. If that is you, fear not! It should run just fine
# in your console.

library(vctrs)

new_nested_iv <- function(iv) {
  fields <- list(iv = iv)
  new_rcrd(fields, class = "nested_iv")
}

format.nested_iv <- function(x, ...) {
  format(field(x, "iv"))
}

iv_proxy.nested_iv <- function(x, ...) {
  field(x, "iv")
}

iv_restore.nested_iv <- function(x, to, ...) {
  new_nested_iv(x)
}

iv <- new_iv(c(1, 5), c(2, 7))

x <- new_nested_iv(iv)
x

# Proxies, then accesses the `start` field
iv_start(x)

# Proxies, computes the complement to generate an iv,
# then restores to the original type
iv_set_complement(x)

}

Group overlapping intervals

Description

This family of functions revolves around grouping overlapping intervals within a single iv. When multiple overlapping intervals are grouped together they result in a wider interval containing the smallest iv_start() and the largest iv_end() of the overlaps.

• iv_groups() merges all overlapping intervals found within x. The resulting intervals are known as the "groups" of x.

• iv_identify_group() identifies the group that the current interval of x falls in. This is particularly useful alongside dplyr::group_by().

• iv_locate_groups() returns a two column data frame with a key column containing the result of iv_groups() and a loc list-column containing integer vectors that map each interval in x to the group that it falls in.

Optionally, you can choose not to group abutting intervals together with abutting = FALSE, which can be useful if you'd like to retain those boundaries.

Minimal interval vectors

iv_groups() is particularly useful because it can generate a minimal interval vector, which covers the range of an interval vector in the most compact form possible. In particular, a minimal interval vector:

• Has no overlapping intervals

• Has no abutting intervals

• Is ordered on both start and end

A minimal interval vector is allowed to have a single missing interval, which is located at the end of the vector.

Usage

iv_groups(x, ..., abutting = TRUE)

iv_identify_group(x, ..., abutting = TRUE)

iv_locate_groups(x, ..., abutting = TRUE)

Arguments

Value

• For iv_groups(), an iv with the same type as x.

• For iv_identify_group(), an iv with the same type and size as x.

• For iv_locate_groups(), a two column data frame with a key column containing the result of iv_groups() and a loc list-column containing integer vectors.

Graphical Representation

Graphically, generating groups looks like:

With abutting = FALSE, intervals that touch aren't grouped:

Examples

library(dplyr, warn.conflicts = FALSE)

x <- iv_pairs(
  c(1, 5),
  c(2, 3),
  c(NA, NA),
  c(5, 6),
  c(NA, NA),
  c(9, 12),
  c(11, 14)
)
x

# Grouping removes all redundancy while still covering the full range
# of values that were originally represented. If any missing intervals
# are present, a single one is retained.
iv_groups(x)

# Abutting intervals are typically grouped together, but you can choose not
# to group them if you want to retain those boundaries
iv_groups(x, abutting = FALSE)

# `iv_identify_group()` is useful alongside `group_by()` and `summarize()`
df <- tibble(x = x)
df <- mutate(df, u = iv_identify_group(x))
df

df %>%
  group_by(u) %>%
  summarize(n = n())

# The real workhorse here is `iv_locate_groups()`, which returns
# the groups and information on which observations in `x` fall in which
# group
iv_locate_groups(x)

Pairwise set operations

Description

This family of functions performs pairwise set operations on two ivs. Pairwise refers to the fact that the i-th interval of x is going to be compared against the i-th interval of y. This is in contrast to their counterparts, like iv_set_union(), which treat the entire vector of x as a single set to be compared against all of y.

The descriptions of these operations are the same as their non-pairwise counterparts, but the ones here also have a number of restrictions due to the fact that each must return an output that is the same size as its inputs:

• For iv_pairwise_set_complement(), x[i] and y[i] can't overlap or abut, as this would generate an empty complement.

• For iv_pairwise_set_union(), x[i] and y[i] can't be separated by a gap. Use iv_pairwise_span() if you want to force gaps to be filled anyways.

• For iv_pairwise_set_intersect(), x[i] and y[i] must overlap, otherwise an empty interval would be generated.

• For iv_pairwise_set_difference(), x[i] can't be completely contained within y[i], as that would generate an empty interval. Additionally, y[i] can't be completely contained within x[i], as that would result in two distinct intervals for a single observation.

• For iv_pairwise_set_symmetric_difference(), x[i] and y[i] must share exactly one endpoint, otherwise an empty interval or two distinct intervals would be generated.

Usage

iv_pairwise_set_complement(x, y)

iv_pairwise_set_union(x, y)

iv_pairwise_set_intersect(x, y)

iv_pairwise_set_difference(x, y)

iv_pairwise_set_symmetric_difference(x, y)

Arguments

Value

An iv the same size and type as x and y.

See Also

The non-pairwise versions of these functions, such as iv_set_union().

Examples

x <- iv_pairs(c(1, 3), c(6, 8))
y <- iv_pairs(c(5, 7), c(2, 3))

iv_pairwise_set_complement(x, y)

z <- iv_pairs(c(2, 5), c(4, 7))

iv_pairwise_set_union(x, z)

# Can't take the union when there are gaps
try(iv_pairwise_set_union(x, y))

# But you can force a union across gaps with `iv_pairwise_span()`
iv_pairwise_span(x, y)

iv_pairwise_set_intersect(x, z)

# Can't take an intersection of non-overlapping intervals
try(iv_pairwise_set_intersect(x, y))

iv_pairwise_set_difference(x, z)

# The pairwise symmetric difference function is fairly strict,
# and is only well defined when exactly one of the interval endpoints match
w <- iv_pairs(c(1, 6), c(7, 8))
iv_pairwise_set_symmetric_difference(x, w)

Pairwise set operations

Description

These functions are deprecated in favor of their set_ prefixed equivalents.

• iv_pairwise_complement() -> iv_pairwise_set_complement()

• iv_pairwise_union() -> iv_pairwise_set_union()

• iv_pairwise_intersect() -> iv_pairwise_set_intersect()

• iv_pairwise_difference() -> iv_pairwise_set_difference()

• iv_pairwise_symmetric_difference() -> iv_pairwise_set_symmetric_difference()

Usage

iv_pairwise_complement(x, y)

iv_pairwise_union(x, y)

iv_pairwise_intersect(x, y)

iv_pairwise_difference(x, y)

iv_pairwise_symmetric_difference(x, y)

Arguments

Set operations

Description

This family of functions treats ivs as sets. They always compute the minimal iv of each input and return a minimal iv.

• iv_set_complement() takes the complement of the intervals in an iv. By default, the minimum and maximum of the inputs define the bounds to take the complement over, but this can be adjusted with lower and upper. Missing intervals are always dropped in the complement.

• iv_set_union() answers the question, "Which intervals are in x or y?" It is equivalent to combining the two vectors together and then calling iv_groups().

• iv_set_intersect() answers the question, "Which intervals are in x and y?"

• iv_set_difference() answers the question, "Which intervals are in x but not y?" Note that this is an asymmetrical difference.

• iv_set_symmetric_difference() answers the question, "Which intervals are in x or y but not both?"

Usage

iv_set_complement(x, ..., lower = NULL, upper = NULL)

iv_set_union(x, y)

iv_set_intersect(x, y)

iv_set_difference(x, y)

iv_set_symmetric_difference(x, y)

Arguments

Value

• For iv_set_complement(), a vector of the same type as x containing the complement.

• For all other set operations, a vector of the same type as the common type of x and y containing the result.

Graphical Representation

Graphically, generating the complement looks like:

If you were to set upper = 20 with these intervals, then you'd get one more interval in the complement.

Generating the intersection between two ivs looks like:

See Also

The pairwise versions of these functions, such as iv_pairwise_set_union().

Examples

x <- iv_pairs(
  c(10, 12),
  c(0, 5),
  c(NA, NA),
  c(3, 6),
  c(-5, -2),
  c(NA, NA)
)
x

y <- iv_pairs(
  c(2, 7),
  c(NA, NA),
  c(-3, -1),
  c(14, 15)
)
y

# Complement contains any values from `[-5, 12)` that aren't represented
# in these intervals. Missing intervals are dropped.
iv_set_complement(x)

# Expand out the "universe" of possible values
iv_set_complement(x, lower = -Inf)
iv_set_complement(x, lower = -Inf, upper = Inf)

# Which intervals are in x or y?
iv_set_union(x, y)

# Which intervals are in x and y?
iv_set_intersect(x, y)

# Which intervals are in x but not y?
iv_set_difference(x, y)

# Which intervals are in y but not x?
iv_set_difference(y, x)

# Missing intervals in x are kept if there aren't missing intervals in y
iv_set_difference(x, iv(1, 2))

# Which intervals are in x or y but not both?
iv_set_symmetric_difference(x, y)

# Missing intervals will be kept if they only appear on one side
iv_set_symmetric_difference(x, iv(1, 2))
iv_set_symmetric_difference(iv(1, 2), x)

Set operations

Description

These functions are deprecated in favor of their set_ prefixed equivalents.

• iv_complement() -> iv_set_complement()

• iv_union() -> iv_set_union()

• iv_intersect() -> iv_set_intersect()

• iv_difference() -> iv_set_difference()

• iv_symmetric_difference() -> iv_set_symmetric_difference()

Usage

iv_complement(x, ..., lower = NULL, upper = NULL)

iv_union(x, y)

iv_intersect(x, y)

iv_difference(x, y)

iv_symmetric_difference(x, y)

Arguments

Splits

Description

This family of functions revolves around splitting an iv on its endpoints, which results in a new iv that is entirely disjoint (i.e. non-overlapping). The intervals in the resulting iv are known as "splits".

• iv_splits() computes the disjoint splits for x.

• iv_identify_splits() identifies the splits that correspond to each interval in x. It replaces x with a list of the same size where each element of the list contains the splits that the corresponding interval in x overlaps. This is particularly useful alongside tidyr::unnest().

• iv_locate_splits() returns a two column data frame with a key column containing the result of iv_splits() and a loc list-column containing integer vectors that map each interval in x to the splits that it overlaps.

Usage

iv_splits(x, ..., on = NULL)

iv_identify_splits(x, ..., on = NULL)

iv_locate_splits(x, ..., on = NULL)

Arguments

Value

• For iv_splits(), an iv with the same type as x.

• For iv_identify_splits(), a list-of containing ivs with the same size as x.

• For iv_locate_splits(), a two column data frame with a key column of the same type as x and loc list-column containing integer vectors.

Graphical Representation

Graphically, generating splits looks like:

Examples

library(tidyr)
library(dplyr)

# Guests to a party and their arrival/departure times
guests <- tibble(
  arrive = as.POSIXct(
    c("2008-05-20 19:30:00", "2008-05-20 20:10:00", "2008-05-20 22:15:00"),
    tz = "UTC"
  ),
  depart = as.POSIXct(
    c("2008-05-20 23:00:00", "2008-05-21 00:00:00", "2008-05-21 00:30:00"),
    tz = "UTC"
  ),
  name = list(
    c("Mary", "Harry"),
    c("Diana", "Susan"),
    "Peter"
  )
)

guests <- unnest(guests, name) %>%
  mutate(iv = iv(arrive, depart), .keep = "unused")

guests

# You can determine the disjoint intervals at which people
# arrived/departed with `iv_splits()`
iv_splits(guests$iv)

# Say you'd like to determine who was at the party at any given time
# throughout the night
guests <- mutate(guests, splits = iv_identify_splits(iv))
guests

# Unnest the splits to generate disjoint intervals for each guest
guests <- guests %>%
  unnest(splits) %>%
  select(name, splits)

guests

# Tabulate who was there at any given time
guests %>%
  summarise(n = n(), who = list(name), .by = splits)

# ---------------------------------------------------------------------------

x <- iv_pairs(c(1, 5), c(4, 9), c(12, 15))
x

# You can provide additional singular values to split on with `on`
iv_splits(x, on = c(2, 13))

Align after locating relationships

Description

iv_align() will align/join needles and haystack together using a data frame of locations. These locations are intended to be the output of one of: iv_locate_overlaps(), iv_locate_precedes(), iv_locate_follows(), iv_locate_relates(), or iv_locate_between().

This is mainly a convenience function that slices both needles and haystack according to those locations, and then stores the result in a new two column data frame.

Usage

iv_align(needles, haystack, ..., locations)

Arguments

Value

A two column data frame with a ⁠$needles⁠ column containing the sliced version of needles and a ⁠$haystack⁠ column containing the sliced version of haystack.

Examples

needles <- iv_pairs(c(1, 5), c(3, 7), c(10, 12))
haystack <- iv_pairs(c(0, 2), c(4, 6))

locations <- iv_locate_overlaps(needles, haystack)
iv_align(needles, haystack, locations = locations)

locations <- iv_locate_overlaps(needles, haystack, no_match = "drop")
iv_align(needles, haystack, locations = locations)

needles <- c(1, 15, 4, 11)
haystack <- iv_pairs(c(1, 5), c(3, 7), c(10, 12))

locations <- iv_locate_between(needles, haystack)
iv_align(needles, haystack, locations = locations)

Diff a vector to create an interval vector

Description

iv_diff() is a convenient way to generate an iv from a preexisting vector, as long as that vector is in strictly increasing order. It returns an iv that is 1 element shorter than x (unless x is already empty).

It is particularly useful for creating an iv column from an existing column inside of dplyr::mutate(), but requires you to explicitly handle padding in that case, see the examples.

Missing values are allowed, and will be propagated to each side of the resulting interval after applying the diff.

Usage

iv_diff(x)

Arguments

Details

iv_diff() is inspired by diff().

Value

An iv using x as the inner type, with size equal to max(0L, vec_size(x) - 1L).

Examples

x <- as.Date("2019-01-01") + c(0, 5, 7, 10, 19)
x

# Notice how the boundaries don't overlap, because the closing `)` aligns
# with an opening `[`.
iv_diff(x)

# Like `iv()`, missing values propagate to both boundaries of the interval.
# Before missing value propagation was applied, it looked like this:
# [1, NA), [NA, 2), [2, 3)
x <- c(1, NA, 2, 3)
iv_diff(x)

# Values in `x` must be in strictly increasing order to generate a valid
# interval vector
x <- c(1, 0, 2, 2)
try(iv_diff(x))

x <- c(1, NA, 0)
try(iv_diff(x))

# ---------------------------------------------------------------------------
# Use with `mutate()`

library(dplyr)

# `iv_diff()` is useful for converting a pre-existing column into an interval
# vector, but you'll need to apply padding to ensure that the size of the
# diff-ed result is the same as the number of rows in your data frame. There
# are two main ways to pad, which are explored below.
df <- tibble(x = c(1, 3, 6))

# Pad with a known lower/upper bound
df %>% mutate(iv = iv_diff(c(0, x)))
df %>% mutate(iv = iv_diff(c(x, Inf)))

# Pad with a missing value, which results in a fully missing interval
df %>% mutate(iv = iv_diff(c(NA, x)))
df %>% mutate(iv = iv_diff(c(x, NA)))

Formatting

Description

iv_format() is an S3 generic intended as a developer tool for making a custom class print nicely when stored in an iv. The default method simply calls format(), and in many cases this is enough for most classes. However, if your class automatically adds justification or padding when formatting a single vector, you might need to implement an iv_format() method to avoid that padding, since it often looks strange when nested in an interval vector.

Usage

iv_format(x)

Arguments

Value

A character vector, likely generated through a call to format().

Examples

# Numeric values get padding automatically through `format()`
x <- c(1, 100)
format(x)

# This ends up looking strange in an iv, so an `iv_format()` method for
# numeric values is implemented which turns off that padding
iv_format(x)

Pairwise span

Description

iv_pairwise_span() computes the pairwise "span" between the i-th interval of x and the i-th interval of y. The pairwise span of two intervals is a new interval containing the minimum start and maximum end of the original intervals. It is similar to iv_pairwise_set_union(), except it fills across gaps.

Usage

iv_pairwise_span(x, y)

Arguments

Value

An iv the same size and type as x and y.

Examples

x <- iv_pairs(c(1, 3), c(6, 8))
y <- iv_pairs(c(5, 7), c(2, 3))

# Can't take the set union when there are gaps
try(iv_pairwise_set_union(x, y))

# But you can compute the span of the intervals
iv_pairwise_span(x, y)

Span

Description

iv_span() computes the span of an iv. The span is a single interval which encompasses the entire range of the iv. It is similar to iv_groups(), if groups were also merged across gaps.

iv_span() is a summary function, like min() and max(), so it always returns a size 1 iv, even for empty ivs. The empty argument can be used to control what is returned in the empty case.

Usage

iv_span(x, ..., missing = "propagate", empty = "missing")

Arguments

Details

iv_span() is currently limited by the fact that it calls min() and max() internally, which doesn't work for all vector types that ivs supports (mainly data frames). In the future, we hope to be able to leverage vctrs::vec_min() and vctrs::vec_max(), which don't exist yet.

Examples

x <- iv_pairs(c(1, 5), c(2, 6), c(9, 10))

# The span covers the full range of values seen in `x`
iv_span(x)

# Compare against `iv_groups()`, which merges overlaps but doesn't merge
# across gaps
iv_groups(x)

x <- iv_pairs(c(1, 3), c(NA, NA), c(5, 6), c(NA, NA))

# Because `iv_span()` is a summary function, if any missing intervals are
# present then it returns a missing interval by default
iv_span(x)

# Further control this with `missing`
iv_span(x, missing = "drop")
try(iv_span(x, missing = "error"))
iv_span(x, missing = iv(-1, 0))

x <- iv(double(), double())

# If `x` is empty, then by default a missing interval is returned
iv_span(x)

# Control this with `empty`
try(iv_span(x, empty = "error"))
iv_span(x, empty = iv(-Inf, Inf))

# `empty` kicks in if `missing = "drop"` is used and all elements were
# missing
x <- iv(c(NA, NA), c(NA, NA), ptype = double())
iv_span(x, missing = "drop", empty = iv(-Inf, Inf))

Construct a new iv

Description

new_iv() is a developer focused function for creating a new interval vector. It does minimal checks on the inputs, for performance.

Usage

new_iv(start, end, ..., class = character())

Arguments

Value

A new iv object.

Examples

new_iv(1, 2)

Count relationships between two ivs

Description

This family of functions counts different types of relationships between two ivs. It works similar to base::match(), where needles[i] checks for a relationship in all of haystack.

• iv_count_overlaps() counts instances of a specific type of overlap between the two ivs.

• iv_count_precedes() counts instances when needles[i] precedes (i.e. comes before) any interval in haystack.

• iv_count_follows() counts instances when needles[i] follows (i.e. comes after) any interval in haystack.

These functions return an integer vector the same size as needles containing a count of the times a particular relationship between the i-th interval of needles and any interval of haystack occurred.

Usage

iv_count_overlaps(
  needles,
  haystack,
  ...,
  type = "any",
  missing = "equals",
  no_match = 0L
)

iv_count_precedes(
  needles,
  haystack,
  ...,
  closest = FALSE,
  missing = "equals",
  no_match = 0L
)

iv_count_follows(
  needles,
  haystack,
  ...,
  closest = FALSE,
  missing = "equals",
  no_match = 0L
)

Arguments

Value

An integer vector the same size as needles.

See Also

Locating relationships

Examples

library(vctrs)

x <- iv_pairs(
  as.Date(c("2019-01-05", "2019-01-10")),
  as.Date(c("2019-01-07", "2019-01-15")),
  as.Date(c("2019-01-20", "2019-01-31"))
)

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Count the number of times `x` overlaps `y` at all
iv_count_overlaps(x, y)

# Count the number of times `y` is within an interval in `x`
iv_count_overlaps(y, x, type = "within")

# Count the number of times `x` precedes `y`
iv_count_precedes(x, y)

# ---------------------------------------------------------------------------

a <- iv(c(1, NA), c(2, NA))
b <- iv(c(NA, NA), c(NA, NA))

# Missing intervals are seen as exactly equal by default, so they are
# considered to overlap
iv_count_overlaps(a, b)

# If you'd like missing intervals to be treated as unmatched, set
# `missing = 0L`
iv_count_overlaps(a, b, missing = 0L)

# If you'd like to propagate missing intervals, set `missing = NA`
iv_count_overlaps(a, b, missing = NA)

Detect a relationship between two ivs

Description

This family of functions detects different types of relationships between two ivs. It works similar to base::%in%, where needles[i] checks for a relationship in all of haystack.

• iv_overlaps() detects a specific type of overlap between the two ivs.

• iv_precedes() detects if needles[i] precedes (i.e. comes before) any interval in haystack.

• iv_follows() detects if needles[i] follows (i.e. comes after) any interval in haystack.

These functions return a logical vector the same size as needles containing TRUE if the interval in needles has a matching relationship in haystack and FALSE otherwise.

Usage

iv_overlaps(needles, haystack, ..., type = "any", missing = "equals")

iv_precedes(needles, haystack, ..., missing = "equals")

iv_follows(needles, haystack, ..., missing = "equals")

Arguments

Value

A logical vector the same size as needles.

See Also

Locating relationships

Detecting relationships pairwise

Locating relations from Allen's Interval Algebra

Examples

library(vctrs)

x <- iv_pairs(
  as.Date(c("2019-01-05", "2019-01-10")),
  as.Date(c("2019-01-07", "2019-01-15")),
  as.Date(c("2019-01-20", "2019-01-31"))
)

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Does each interval of `x` overlap `y` at all?
iv_overlaps(x, y)

# Which intervals of `y` are within an interval in `x`?
iv_overlaps(y, x, type = "within")

# ---------------------------------------------------------------------------

a <- iv(c(1, NA), c(2, NA))
b <- iv(c(NA, NA), c(NA, NA))

# Missing intervals are seen as exactly equal by default, so they are
# considered to overlap
iv_overlaps(a, b)

# If you'd like missing intervals to be treated as unmatched, set
# `missing = FALSE`
iv_overlaps(a, b, missing = FALSE)

# If you'd like to propagate missing intervals, set `missing = NA`
iv_overlaps(a, b, missing = NA)

Pairwise detect a relationship between two ivs

Description

This family of functions detects different types of relationships between two ivs pairwise, where pairwise means that the i-th interval of x is compared against the i-th interval of y. This is in contrast to iv_overlaps(), which works more like base::%in%.

• iv_pairwise_overlaps() detects a specific type of overlap between the i-th interval of x and the i-th interval of y.

• iv_pairwise_precedes() detects if the i-th interval of x precedes (i.e. comes before) the i-th interval of y.

• iv_pairwise_follows() detects if the i-th interval of x follows (i.e. comes after) the i-th interval of y.

These functions return a logical vector the same size as the common size of x and y.

Usage

iv_pairwise_overlaps(x, y, ..., type = "any")

iv_pairwise_precedes(x, y)

iv_pairwise_follows(x, y)

Arguments

Value

A logical vector the same size as the common size of x and y.

See Also

Locating relationships

Detecting relationships

Locating relations from Allen's Interval Algebra

Examples

library(vctrs)

x <- iv_pairs(
  as.Date(c("2019-01-05", "2019-01-10")),
  as.Date(c("2019-01-07", "2019-01-15")),
  as.Date(c("2019-01-20", "2019-01-31"))
)

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-18", "2019-01-21"))
)

x
y

# Does the i-th interval of `x` overlap the i-th interval of `y`?
iv_pairwise_overlaps(x, y)

# Does the i-th interval of `x` contain the i-th interval of `y`?
iv_pairwise_overlaps(x, y, type = "contains")

# Does the i-th interval of `x` follow the i-th interval of `y`?
iv_pairwise_follows(x, y)

a <- iv_pairs(c(1, 2), c(NA, NA), c(NA, NA))
b <- iv_pairs(c(NA, NA), c(3, 4), c(NA, NA))

# Missing intervals always propagate
iv_pairwise_overlaps(a, b)

Locate relationships between two ivs

Description

This family of functions locates different types of relationships between two ivs. It works similar to base::match(), where needles[i] checks for a relationship in all of haystack. Unlike match(), all matching relationships are returned, rather than just the first.

• iv_locate_overlaps() locates a specific type of overlap between the two ivs.

• iv_locate_precedes() locates where needles[i] precedes (i.e. comes before) any interval in haystack.

• iv_locate_follows() locates where needles[i] follows (i.e. comes after) any interval in haystack.

These functions return a two column data frame. The needles column is an integer vector pointing to locations in needles. The haystack column is an integer vector pointing to locations in haystack with a matching relationship.

Usage

iv_locate_overlaps(
  needles,
  haystack,
  ...,
  type = "any",
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

iv_locate_precedes(
  needles,
  haystack,
  ...,
  closest = FALSE,
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

iv_locate_follows(
  needles,
  haystack,
  ...,
  closest = FALSE,
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

Arguments

Value

A data frame containing two integer columns named needles and haystack.

See Also

Detecting relationships

Detecting relationships pairwise

Locating relations from Allen's Interval Algebra

Examples

x <- iv_pairs(
  as.Date(c("2019-01-05", "2019-01-10")),
  as.Date(c("2019-01-07", "2019-01-15")),
  as.Date(c("2019-01-20", "2019-01-31"))
)

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Find any overlap between `x` and `y`
loc <- iv_locate_overlaps(x, y)
loc

iv_align(x, y, locations = loc)

# Find where `x` contains `y` and drop results when there isn't a match
loc <- iv_locate_overlaps(x, y, type = "contains", no_match = "drop")
loc

iv_align(x, y, locations = loc)

# Find where `x` precedes `y`
loc <- iv_locate_precedes(x, y)
loc

iv_align(x, y, locations = loc)

# Filter down to find only the closest interval in `y` of all the intervals
# where `x` preceded it
loc <- iv_locate_precedes(x, y, closest = TRUE)

iv_align(x, y, locations = loc)

# Note that `closest` can result in duplicates if there is a tie.
# `2019-01-20` appears as an end date twice in `haystack`.
loc <- iv_locate_follows(x, y, closest = TRUE)
loc

iv_align(x, y, locations = loc)

# Force just one of the ties to be returned by using `multiple`.
# Here we just request any of the ties, with no guarantee on which one.
loc <- iv_locate_follows(x, y, closest = TRUE, multiple = "any")
loc

iv_align(x, y, locations = loc)

# ---------------------------------------------------------------------------

a <- iv(NA, NA)
b <- iv(c(NA, NA), c(NA, NA))

# By default, missing intervals in `needles` are seen as exactly equal to
# missing intervals in `haystack`, which means that they overlap
iv_locate_overlaps(a, b)

# If you'd like missing intervals in `needles` to always be considered
# unmatched, set `missing = NA`
iv_locate_overlaps(a, b, missing = NA)

Count relationships between a vector and an iv

Description

This family of functions counts different types of relationships between a vector and an iv. It works similar to base::match(), where needles[i] checks for a match in all of haystack.

• iv_count_between() counts instances of when needles, a vector, falls between the bounds of haystack, an iv.

• iv_count_includes() counts instances of when needles, an iv, includes the values of haystack, a vector.

These functions return an integer vector the same size as needles containing a count of the times where the i-th value of needles contained a match in haystack.

Usage

iv_count_between(needles, haystack, ..., missing = "equals", no_match = 0L)

iv_count_includes(needles, haystack, ..., missing = "equals", no_match = 0L)

Arguments

Value

An integer vector the same size as needles.

See Also

Locating relationships between a vector and an iv

Examples

x <- as.Date(c("2019-01-05", "2019-01-10", "2019-01-07", "2019-01-20"))

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Count the number of times `x` is between the intervals in `y`
iv_count_between(x, y)

# Count the number of times `y` includes a value from `x`
iv_count_includes(y, x)

# ---------------------------------------------------------------------------

a <- c(1, NA)
b <- iv(c(NA, NA), c(NA, NA))

# By default, missing values in `needles` are treated as being exactly
# equal to missing values in `haystack`, so the missing value in `a` is
# considered between the missing interval in `b`.
iv_count_between(a, b)
iv_count_includes(b, a)

# If you'd like to propagate missing values, set `missing = NA`
iv_count_between(a, b, missing = NA)
iv_count_includes(b, a, missing = NA)

# If you'd like missing values to be treated as unmatched, set
# `missing = 0L`
iv_count_between(a, b, missing = 0L)
iv_count_includes(b, a, missing = 0L)

Detect relationships between a vector and an iv

Description

This family of functions detects different types of relationships between a vector and an iv. It works similar to base::%in%, where needles[i] checks for a match in all of haystack.

• iv_between() detects when needles, a vector, falls between the bounds in haystack, an iv.

• iv_includes() detects when needles, an iv, includes the values of haystack, a vector.

This function returns a logical vector the same size as needles containing TRUE if the value in needles matches any value in haystack and FALSE otherwise.

Usage

iv_between(needles, haystack, ..., missing = "equals")

iv_includes(needles, haystack, ..., missing = "equals")

Arguments

Value

A logical vector the same size as needles.

See Also

Locating relationships

Locating relationships between a vector and an iv

Pairwise detect relationships between a vector and an iv

Examples

x <- as.Date(c("2019-01-05", "2019-01-10", "2019-01-07", "2019-01-20"))

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Detect if the i-th location in `x` is between any intervals in `y`
iv_between(x, y)

# Detect if the i-th location in `y` includes any value in `x`
iv_includes(y, x)

# ---------------------------------------------------------------------------

a <- c(1, NA)
b <- iv(c(NA, NA), c(NA, NA))

# By default, missing values in `needles` are treated as being exactly
# equal to missing values in `haystack`, so the missing value in `a` is
# considered between the missing interval in `b`.
iv_between(a, b)
iv_includes(b, a)

# If you'd like to propagate missing values, set `missing = NA`
iv_between(a, b, missing = NA)
iv_includes(b, a, missing = NA)

# If you'd like missing values to be treated as unmatched, set
# `missing = FALSE`
iv_between(a, b, missing = FALSE)
iv_includes(b, a, missing = FALSE)

Pairwise detect relationships between a vector and an iv

Description

This family of functions detects different types of relationships between a vector and an iv pairwise. where pairwise means that the i-th value of x is compared against the i-th value of y. This is in contrast to iv_between(), which works more like base::%in%.

• iv_pairwise_between() detects if the i-th value of x, a vector, falls between the bounds of the i-th value of y, an iv.

• iv_pairwise_includes() detects if the i-th value of x, an iv, includes the i-th value of y, a vector.

These functions return a logical vector the same size as the common size of x and y.

Usage

iv_pairwise_between(x, y)

iv_pairwise_includes(x, y)

Arguments

Value

A logical vector the same size as the common size of x and y.

See Also

Locating relationships

Locating relationships between a vector and an iv

Detecting relationships between a vector and an iv

Examples

x <- as.Date(c("2019-01-01", "2019-01-08", "2019-01-21"))

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-18", "2019-01-21"))
)

x
y

# Does the i-th value of `x` fall between the i-th interval of `y`?
iv_pairwise_between(x, y)

# Does the i-th interval of `y` include the i-th value of `x`?
iv_pairwise_includes(y, x)

a <- c(1, NA, NA)
b <- iv_pairs(c(NA, NA), c(3, 4), c(NA, NA))

# Missing intervals always propagate
iv_pairwise_between(a, b)
iv_pairwise_includes(b, a)

Locate relationships between a vector and an iv

Description

This family of functions locates different types of relationships between a vector and an iv. It works similar to base::match(), where needles[i] checks for a match in all of haystack. Unlike match(), all matches are returned, rather than just the first.

• iv_locate_between() locates where needles, a vector, falls between the bounds of haystack, an iv.

• iv_locate_includes() locates where needles, an iv, includes the values of haystack, a vector.

These functions return a two column data frame. The needles column is an integer vector pointing to locations in needles. The haystack column is an integer vector pointing to locations in haystack with a match.

Usage

iv_locate_between(
  needles,
  haystack,
  ...,
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

iv_locate_includes(
  needles,
  haystack,
  ...,
  missing = "equals",
  no_match = NA_integer_,
  remaining = "drop",
  multiple = "all",
  relationship = "none"
)

Arguments

Value

A data frame containing two integer columns named needles and haystack.

See Also

Locating relationships

Detect relationships between a vector and an iv

Pairwise detect relationships between a vector and an iv

Examples

x <- as.Date(c("2019-01-05", "2019-01-10", "2019-01-07", "2019-01-20"))

y <- iv_pairs(
  as.Date(c("2019-01-01", "2019-01-03")),
  as.Date(c("2019-01-04", "2019-01-08")),
  as.Date(c("2019-01-07", "2019-01-09")),
  as.Date(c("2019-01-10", "2019-01-20")),
  as.Date(c("2019-01-15", "2019-01-20"))
)

x
y

# Find any location where `x` is between the intervals in `y`
loc <- iv_locate_between(x, y)
loc

iv_align(x, y, locations = loc)

# Find any location where `y` includes the values in `x`
loc <- iv_locate_includes(y, x)
loc

iv_align(y, x, locations = loc)

# Drop values in `x` without a match
loc <- iv_locate_between(x, y, no_match = "drop")
loc

iv_align(x, y, locations = loc)

# ---------------------------------------------------------------------------

a <- c(1, NA)
b <- iv(c(NA, NA), c(NA, NA))

# By default, missing values in `needles` are treated as being exactly
# equal to missing values in `haystack`, so the missing value in `a` is
# considered between the missing interval in `b`.
iv_locate_between(a, b)
iv_locate_includes(b, a)

# If you'd like missing values in `needles` to always be considered
# unmatched, set `missing = NA`
iv_locate_between(a, b, missing = NA)
iv_locate_includes(b, a, missing = NA)