Improve sort!, searchsortedXXX and insorted doc

Author: knuesel

base/operators.jl

@@ -163,8 +163,6 @@ Types with a partial order should implement [`<`](@ref).
 See the documentation on [Alternate Orderings](@ref) for how to define alternate
 ordering methods that can be used in sorting and related functions.
 
-See also [`isequal`](@ref), [`isunordered`](@ref).
-
 # Examples
 ```jldoctest
 julia> isless(1, 3)
@@ -330,6 +328,8 @@ New types with a canonical partial order should implement this function for
 two arguments of the new type.
 Types with a canonical total order should implement [`isless`](@ref) instead.
 
+See also [`isunordered`](@ref).
+
 # Examples
 ```jldoctest
 julia> 'a' < 'b'

base/sort.jl

@@ -162,8 +162,9 @@ partialsort(v::AbstractVector, k::Union{Integer,OrdinalRange}; kws...) =
 # reference on sorted binary search:
 #   http://www.tbray.org/ongoing/When/200x/2003/03/22/Binary
 
-# index of the first value of vector a that is greater than or equal to x;
-# returns lastindex(v)+1 if x is greater than all values in v.
+# Index of the first value of vector a that doesn't sort before x (for the
+# default order this means the first value less than or equivalent to x).
+# Returns lastindex(v)+1 if x sorts after all values of v.
 function searchsortedfirst(v::AbstractVector, x, lo::T, hi::T, o::Ordering)::keytype(v) where T<:Integer
     hi = hi + T(1)
     len = hi - lo
@@ -181,8 +182,9 @@ function searchsortedfirst(v::AbstractVector, x, lo::T, hi::T, o::Ordering)::key
     return lo
 end
 
-# index of the last value of vector a that is less than or equal to x;
-# returns firstindex(v)-1 if x is less than all values of v.
+# Index of the last value of vector a that doesn't sort after x (for the
+# default order this means the last value less than or equaivalent to x).
+# returns firstindex(v)-1 if x sorts before all values of v.
 function searchsortedlast(v::AbstractVector, x, lo::T, hi::T, o::Ordering)::keytype(v) where T<:Integer
     u = T(1)
     lo = lo - u
@@ -198,7 +200,7 @@ function searchsortedlast(v::AbstractVector, x, lo::T, hi::T, o::Ordering)::keyt
     return lo
 end
 
-# returns the range of indices of v equal to x
+# returns the range of indices of v equivalent to x
 # if v does not contain x, returns a 0-length range
 # indicating the insertion point of x
 function searchsorted(v::AbstractVector, x, ilo::T, ihi::T, o::Ordering)::UnitRange{keytype(v)} where T<:Integer
@@ -291,16 +293,24 @@ for s in [:searchsortedfirst, :searchsortedlast, :searchsorted]
 end
 
 """
-    searchsorted(a, x; by=identity, lt=isless, rev=false)
+    searchsorted(v, x; by=identity, lt=isless, rev=false)
 
-Return the range of indices of `a` that compare as equal to `x` (using binary
-search), assuming that `a` is already sorted. Return an empty range located at
-the insertion point if `a` does not contain values equal to `x`.
+Return the range of indices in `v` where values are equivalent to `x`, or an
+empty range located at the insertion point if `v` does not contain values
+equivalent to `x` (see below for the definition of equivalence).
+
+The range is found using binary search. The vector `v` must be sorted, or at
+least partitioned with respect to `x` such that all values that sort before `x`
+come first and all values that sort after `x` come last.
 
 The `by`, `lt` and `rev` keywords modify what order is assumed for the data,
 as described in the [`sort!`](@ref) documentation.
 
-See also: [`insorted`](@ref), [`searchsortedfirst`](@ref), [`sort!`](@ref), [`findall`](@ref).
+Values `x` and `y` are said equivalent if `lt(by(x), by(y))` and `lt(by(y),
+by(x))` both return `false`, and `x` is said to sort before `y` if `lt(by(x),
+by(y))` returns `true` (with `x` and `y` exchanged if `rev=true`).
+
+See also: [`searchsortedfirst`](@ref), [`sort!`](@ref), [`insorted`](@ref), [`findall`](@ref).
 
 # Examples
 ```jldoctest
@@ -318,20 +328,30 @@ julia> searchsorted([1, 2, 4, 5, 5, 7], 9) # no match, insert at end
 
 julia> searchsorted([1, 2, 4, 5, 5, 7], 0) # no match, insert at start
 1:0
+
+julia> searchsorted([1, 0, 0, 2, 2, 7, 6], 2) # data unsorted but partitioned with respect to 2
+4:5
+
+julia> searchsorted([1, -1, -2, 2, -2, 3, -4, 4], 2, by=abs) # sorted by absolute value, -2 equivalent to 2
+3:5
 ```
 """ searchsorted
 
 """
-    searchsortedfirst(a, x; by=identity, lt=isless, rev=false)
+    searchsortedfirst(v, x; by=identity, lt=isless, rev=false)
 
-Return the index of the first value in `a` greater than or equal to `x`,
-assuming that `a` is already sorted. Return `lastindex(a) + 1` if `x` is
-greater than all values in `a`.
+Return the index of the first value in `v` that doesn't sort before `x`.
+If all values in `v` sort before `x`, the function returns `lastindex(v) + 1`.
 
-`insert!`ing `x` at this index will maintain sorted order.
+The index is found using binary search. The vector `v` must be sorted, or at
+least partitioned with respect to `x` such that all values that sort before `x`
+come first and all values that sort after `x` come last. `insert!`ing `x` at
+the returned index will maintain the sorted order (or partition).
 
 The `by`, `lt` and `rev` keywords modify what order is assumed for the data,
-as described in the [`sort!`](@ref) documentation.
+as described in the [`sort!`](@ref) documentation. Value `x` is said to sort
+before `y` if `lt(by(x), by(y))` returns `true` (with `x` and `y` exchanged if
+`rev=true`).
 
 See also: [`searchsortedlast`](@ref), [`searchsorted`](@ref), [`findfirst`](@ref).
 
@@ -351,16 +371,29 @@ julia> searchsortedfirst([1, 2, 4, 5, 5, 7], 9) # no match, insert at end
 
 julia> searchsortedfirst([1, 2, 4, 5, 5, 7], 0) # no match, insert at start
 1
+
+julia> searchsortedfirst([1, 0, 0, 2, 2, 7, 6], 2) # data unsorted but partitioned with respect to 2
+4
+
+julia> searchsortedfirst([1, -1, -2, 2, -2, 3, -4, 4], 2, by=abs) # sorted by absolute value
+3
 ```
 """ searchsortedfirst
 
 """
-    searchsortedlast(a, x; by=identity, lt=isless, rev=false)
+    searchsortedlast(v, x; by=identity, lt=isless, rev=false)
 
-Return the index of the last value in `a` less than or equal to `x`, assuming
-that `a` is already sorted. Return `firstindex(a) - 1` if `x` is less than all
-values in `a`. The `by`, `lt` and `rev` keywords modify what order is assumed
-for the data, as described in the [`sort!`](@ref) documentation.
+Return the index of the last value in `v` that doesn't sort after `x`.
+If all values in `v` sort after `x`, the function returns `firstindex(v) - 1`.
+
+The index is found using binary search. The vector `v` must be sorted, or at
+least partitioned with respect to `x` such that all values that sort before `x`
+come first and all values that sort after `x` come last.
+
+The `by`, `lt` and `rev` keywords modify what order is assumed for the data,
+as described in the [`sort!`](@ref) documentation. Value `x` is said to sort
+before `y` if `lt(by(x), by(y))` returns `true` (with `x` and `y` exchanged if
+`rev=true`).
 
 # Examples
 ```jldoctest
@@ -378,16 +411,31 @@ julia> searchsortedlast([1, 2, 4, 5, 5, 7], 9) # no match, insert at end
 
 julia> searchsortedlast([1, 2, 4, 5, 5, 7], 0) # no match, insert at start
 0
+
+julia> searchsortedlast([1, 0, 0, 2, 2, 7, 6], 2) # data unsorted but partitioned with respect to 2
+5
+
+julia> searchsortedlast([1, -1, -2, 2, -2, 3, -4, 4], 2, by=abs) # sorted by absolute value
+5
 ```
 """ searchsortedlast
 
 """
-    insorted(x, a; by=identity, lt=isless, rev=false) -> Bool
+    insorted(x, v; by=identity, lt=isless, rev=false) -> Bool
 
-Determine whether an item `x` is in the sorted collection `a`, in the sense that
-it is [`==`](@ref) to one of the values of the collection. The `by`, `lt` and
-`rev` keywords modify what order is assumed for the collection, as described in
-the [`sort!`](@ref) documentation.
+Determine whether a vector `v` contains any value equivalent to `x` (see below
+for the definition of equivalence).
+
+The check is done using binary search. The vector `v` must be sorted, or at
+least partitioned with respect to `x` such that all values that sort before `x`
+come first and all values that sort after `x` come last.
+
+The `by`, `lt` and `rev` keywords modify what order is assumed for the data,
+as described in the [`sort!`](@ref) documentation.
+
+Values `x` and `y` are said equivalent if `lt(by(x), by(y))` and `lt(by(y),
+by(x))` both return `false`, and `x` is said to sort before `y` if `lt(by(x),
+by(y))` returns `true` (with `x` and `y` exchanged if `rev=true`).
 
 See also [`in`](@ref).
 
@@ -407,6 +455,12 @@ false
 
 julia> insorted(0, [1, 2, 4, 5, 5, 7]) # no match
 false
+
+julia> insorted(2, [1, 0, 2, 2, 7, 6]) # data unsorted but partitioned with respect to 2
+true
+
+julia> insorted(2, [1, -1, -2, 3, -4, 4], by=abs) # sorted by absolute value
+true
 ```
 
 !!! compat "Julia 1.6"
@@ -1338,11 +1392,14 @@ for available algorithms). Elements are first transformed by the function `by`
 and then compared according to either the function `lt` or the ordering
 `order`. Finally, the resulting order is reversed if `rev=true`.
 
-The `lt` function should define a strict partial order, that is, it should be
+The `lt` function should define a strict weak ordering, that is, it should be
 
 - irreflexive: `lt(x, x)` always yields `false`,
 - asymmetric: if `lt(x, y)` yields `true` then `lt(y, x)` yields `false`,
-- transitive: `lt(x, y) && lt(y, z)` implies `lt(x, z)`.
+- transitive: `lt(x, y) && lt(y, z)` implies `lt(x, z)`,
+- transitive in incomparability: `incomparable(x, y) && incomparable(y, z)`
+  implies `incomparable(x, z)`, where `incomparable(x, y)` is defined as
+  `!lt(x, y) && !lt(y, x)`.
 
 For example `<` is a valid `lt` function but `≤` is not.