Struct TimestampDifference

struct TimestampDifference { ... }

Options for Timestamp::since and Timestamp::until.

This type provides a way to configure the calculation of spans between two Timestamp values. In particular, both Timestamp::since and Timestamp::until accept anything that implements Into<TimestampDifference>. There are a few key trait implementations that make this convenient:

One can also provide a TimestampDifference value directly. Doing so is necessary to use the rounding features of calculating a span. For example, setting the smallest unit (defaults to Unit::Nanosecond), the rounding mode (defaults to RoundMode::Trunc) and the rounding increment (defaults to 1). The defaults are selected such that no rounding occurs.

Rounding a span as part of calculating it is provided as a convenience. Callers may choose to round the span as a distinct step via Span::round.

Example

This example shows how to round a span between two timestamps to the nearest half-hour, with ties breaking away from zero.

use jiff::{RoundMode, Timestamp, TimestampDifference, ToSpan, Unit};

let ts1 = "2024-03-15 08:14:00.123456789Z".parse::<Timestamp>()?;
let ts2 = "2024-03-22 15:00Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Minute)
        .largest(Unit::Hour)
        .mode(RoundMode::HalfExpand)
        .increment(30),
)?;
assert_eq!(format!("{span:#}"), "175h");

// One less minute, and because of the HalfExpand mode, the span would
// get rounded down.
let ts2 = "2024-03-22 14:59Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Minute)
        .largest(Unit::Hour)
        .mode(RoundMode::HalfExpand)
        .increment(30),
)?;
assert_eq!(span, 174.hours().minutes(30).fieldwise());

# Ok::<(), Box<dyn std::error::Error>>(())

Implementations

impl TimestampDifference

fn new(timestamp: Timestamp) -> TimestampDifference

Create a new default configuration for computing the span between the given timestamp and some other time (specified as the receiver in Timestamp::since or Timestamp::until).

fn smallest(self: Self, unit: Unit) -> TimestampDifference

Set the smallest units allowed in the span returned.

Errors

The smallest units must be no greater than the largest units. If this is violated, then computing a span with this configuration will result in an error.

Example

This shows how to round a span between two timestamps to units no less than seconds.

use jiff::{RoundMode, Timestamp, TimestampDifference, ToSpan, Unit};

let ts1 = "2024-03-15 08:14:02.5001Z".parse::<Timestamp>()?;
let ts2 = "2024-03-15T08:16:03.0001Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Second)
        .mode(RoundMode::HalfExpand),
)?;
assert_eq!(span, 121.seconds().fieldwise());

// Because of the rounding mode, a small less-than-1-second increase in
// the first timestamp can change the result of rounding.
let ts1 = "2024-03-15 08:14:02.5002Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Second)
        .mode(RoundMode::HalfExpand),
)?;
assert_eq!(span, 120.seconds().fieldwise());

# Ok::<(), Box<dyn std::error::Error>>(())
fn largest(self: Self, unit: Unit) -> TimestampDifference

Set the largest units allowed in the span returned.

When a largest unit is not specified, computing a span between timestamps behaves as if it were set to Unit::Second. Unless TimestampDifference::smallest is bigger than Unit::Second, then the largest unit is set to the smallest unit.

Errors

The largest units, when set, must be at least as big as the smallest units (which defaults to Unit::Nanosecond). If this is violated, then computing a span with this configuration will result in an error.

Example

This shows how to round a span between two timestamps to units no bigger than seconds.

use jiff::{Timestamp, TimestampDifference, ToSpan, Unit};

let ts1 = "2024-03-15 08:14Z".parse::<Timestamp>()?;
let ts2 = "2030-11-22 08:30Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2).largest(Unit::Second),
)?;
assert_eq!(format!("{span:#}"), "211076160s");

# Ok::<(), Box<dyn std::error::Error>>(())
fn mode(self: Self, mode: RoundMode) -> TimestampDifference

Set the rounding mode.

This defaults to RoundMode::Trunc since it's plausible that rounding "up" in the context of computing the span between two timestamps could be surprising in a number of cases. The RoundMode::HalfExpand mode corresponds to typical rounding you might have learned about in school. But a variety of other rounding modes exist.

Example

This shows how to always round "up" towards positive infinity.

use jiff::{RoundMode, Timestamp, TimestampDifference, ToSpan, Unit};

let ts1 = "2024-03-15 08:10Z".parse::<Timestamp>()?;
let ts2 = "2024-03-15 08:11Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Hour)
        .mode(RoundMode::Ceil),
)?;
// Only one minute elapsed, but we asked to always round up!
assert_eq!(span, 1.hour().fieldwise());

// Since `Ceil` always rounds toward positive infinity, the behavior
// flips for a negative span.
let span = ts1.since(
    TimestampDifference::new(ts2)
        .smallest(Unit::Hour)
        .mode(RoundMode::Ceil),
)?;
assert_eq!(span, 0.hour().fieldwise());

# Ok::<(), Box<dyn std::error::Error>>(())
fn increment(self: Self, increment: i64) -> TimestampDifference

Set the rounding increment for the smallest unit.

The default value is 1. Other values permit rounding the smallest unit to the nearest integer increment specified. For example, if the smallest unit is set to Unit::Minute, then a rounding increment of 30 would result in rounding in increments of a half hour. That is, the only minute value that could result would be 0 or 30.

Errors

The rounding increment must divide evenly into the next highest unit after the smallest unit configured (and must not be equivalent to it). For example, if the smallest unit is Unit::Nanosecond, then some of the valid values for the rounding increment are 1, 2, 4, 5, 100 and 500. Namely, any integer that divides evenly into 1,000 nanoseconds since there are 1,000 nanoseconds in the next highest unit (microseconds).

The error will occur when computing the span, and not when setting the increment here.

Example

This shows how to round the span between two timestamps to the nearest 5 minute increment.

use jiff::{RoundMode, Timestamp, TimestampDifference, ToSpan, Unit};

let ts1 = "2024-03-15 08:19Z".parse::<Timestamp>()?;
let ts2 = "2024-03-15 12:52Z".parse::<Timestamp>()?;
let span = ts1.until(
    TimestampDifference::new(ts2)
        .smallest(Unit::Minute)
        .increment(5)
        .mode(RoundMode::HalfExpand),
)?;
assert_eq!(span.to_string(), "PT275M");

# Ok::<(), Box<dyn std::error::Error>>(())

impl Clone for TimestampDifference

fn clone(self: &Self) -> TimestampDifference

impl Copy for TimestampDifference

impl Debug for TimestampDifference

fn fmt(self: &Self, f: &mut Formatter<'_>) -> Result

impl Freeze for TimestampDifference

impl From for TimestampDifference

fn from(zdt: Zoned) -> TimestampDifference

impl From for TimestampDifference

fn from((largest, ts): (Unit, Timestamp)) -> TimestampDifference

impl From for TimestampDifference

fn from((largest, zdt): (Unit, Zoned)) -> TimestampDifference

impl From for TimestampDifference

fn from(ts: Timestamp) -> TimestampDifference

impl RefUnwindSafe for TimestampDifference

impl Send for TimestampDifference

impl Sync for TimestampDifference

impl Unpin for TimestampDifference

impl UnsafeUnpin for TimestampDifference

impl UnwindSafe for TimestampDifference

impl<'a> From for TimestampDifference

fn from(zdt: &'a Zoned) -> TimestampDifference

impl<'a> From for TimestampDifference

fn from((largest, zdt): (Unit, &'a Zoned)) -> TimestampDifference

impl<T> Any for TimestampDifference

fn type_id(self: &Self) -> TypeId

impl<T> Borrow for TimestampDifference

fn borrow(self: &Self) -> &T

impl<T> BorrowMut for TimestampDifference

fn borrow_mut(self: &mut Self) -> &mut T

impl<T> CloneToUninit for TimestampDifference

unsafe fn clone_to_uninit(self: &Self, dest: *mut u8)

impl<T> From for TimestampDifference

fn from(t: T) -> T

Returns the argument unchanged.

impl<T> ToOwned for TimestampDifference

fn to_owned(self: &Self) -> T
fn clone_into(self: &Self, target: &mut T)

impl<T, U> Into for TimestampDifference

fn into(self: Self) -> U

Calls U::from(self).

That is, this conversion is whatever the implementation of [From]<T> for U chooses to do.

impl<T, U> TryFrom for TimestampDifference

fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>

impl<T, U> TryInto for TimestampDifference

fn try_into(self: Self) -> Result<U, <U as TryFrom<T>>::Error>