An alternative interface for dictionaries in Julia, for improved productivity and performance
This package is somewhat young - new features are being added and some (low-level) interfaces may be tweaked in the future, but things should be stable enough for general usage. Contributions welcome - please submit an issue or PR!
The high-level goal of this package is to define a new interface for dictionary and set structures which is convenient and efficient for functional data manipulation - including operations such as non-scalar indexing, broadcasting, mapping, filtering, reducing, grouping, and so-on. While Julia comes with built-in AbstractDict and AbstractSet supertypes, the interfaces for these are not as well established or generic as for AbstractArray, the built-in dictionaries implement less of the common data manipulation operations compared to arrays, and it is difficult to work with them in a performant manner.
In this package we aim to devise a cohesive interface for abstract dictionaries (or associative maps), having the common supertype AbstractDictionary. A large part of this is working with indices (of arbitrary type) as well as convenient and efficient iteration of the containers. A second goal is to make dictionary manipulation more closely resemble array manipulation, to make it easier for users. Simultaneously, we are pushing the performance of working with dictionaries to be closer to that of working with arrays.
Dictionaries share the common supertype AbstractDictionary, and the go-to container in this package is Dictionary - which is a new hash-based implementation that serves as a replacement of Julia's inbuilt Dict type (using hash and isequal for key lookup and comparison). The three main difference to Dict are that it preserves the order of elements, it iterates much faster, and it iterates values rather than key-value pairs.
You can construct one from a list of indices (or keys) and a list of values.
julia> dict = Dictionary(["a", "b", "c"], [1, 2, 3])
3-element Dictionary{String,Int64}
"a" │ 1
"b" │ 2
"c" │ 3
julia> dict["a"]
1
julia> dict == Dictionary(("a", "b", "c"), (1, 2, 3))
trueThe constructor also accepts any indexable container, preserving the keys and values.
julia> Dictionary(Dict("a"=>1, "b"=>2, "c"=>3))
3-element Dictionary{String,Int64}
"c" │ 3
"b" │ 2
"a" │ 1If you prefer, you can use the dictionary function to create a dictionary from something
that iterates key-value pairs (either as a Pair or a two-tuple, etc), somewhat like a
Dict constructor.
julia> dictionary(["a" => 1, "b" => 2, "c" => 3])
3-element Dictionary{String,Int64}
"a" │ 1
"b" │ 2
"c" │ 3One final way to construct a dictionary is using the index function, which accepts a
function that constructs a "key" for each element in the collection.
julia> index(first, ["Alice", "Bob", "Charlie"])
3-element Dictionary{Char,String}
'A' │ "Alice"
'B' │ "Bob"
'C' │ "Charlie"The values of Dictionary are mutable, or "settable", and can be modified via setindex!.
However, just like for Arrays, new indices (keys) are never created or rearranged this way.
julia> dict["a"] = 10
10
julia> dict
3-element Dictionary{String,Int64}
"a" │ 10
"b" │ 2
"c" │ 3
julia> dict["d"] = 42
ERROR: IndexError("Dictionary does not contain index: d")
Stacktrace:
[1] setindex!(::Dictionary{String,Int64}, ::Int64, ::String) at /home/ferris/.julia/dev/Dictionaries/src/AbstractDictionary.jl:347
[2] top-level scope at REPL[7]:1The indices of Dictionary are said to be "insertable" - indices can be added or removed with the insert! and delete! functions.
julia> insert!(dict, "d", 42)
4-element Dictionary{String,Int64}
"a" │ 10
"b" │ 2
"c" │ 3
"d" │ 42
julia> delete!(dict, "d")
3-element Dictionary{String,Int64}
"a" │ 10
"b" │ 2
"c" │ 3
Note that insert! and delete! are precise in the sense that insert! will error if the index already exists, and delete! will error if the index does not. The set! function provides "upsert" functionality ("update or insert") and unset! is useful for removing an index that may or may not exist.
Dictionaries can be manipulated and transformed using a similar interface to Julia's built-in arrays. The first thing to note is that dictionaries iterate values, so it easy to perform simple analytics on the dictionary values.
julia> dict = Dictionary(["a", "b", "c"], [1, 2, 3])
3-element Dictionary{String,Int64}
"a" │ 1
"b" │ 2
"c" │ 3
julia> sum(dict)
6
julia> using Statistics; mean(dict)
2.0Mapping and broadcasting also function as-per arrays, preserving the indices and transforming the corresponding values.
julia> map(iseven, dict)
3-element Dictionary{String,Bool}
"a" │ false
"b" │ true
"c" │ false
julia> map(*, dict, dict)
3-element Dictionary{String,Int64}
"a" │ 1
"b" │ 4
"c" │ 9
julia> dict .+ 1
3-element Dictionary{String,Int64}
"a" │ 2
"b" │ 3
"c" │ 4There is a mapview function exported by SplitApplyCombine.jl, which is the lazy version of the above.
Filtering a dictionary also preserves the keys, dropping the remainder.
julia> filter(isodd, dict)
2-element Dictionary{String,Bool}
"a" │ 1
"c" │ 3The filterview function is provided to lazily filter a dictionary, which may occasionally
be more performant when working with larger containers.
The pairs function allows access to both the index (key) and value when iterating.
julia> pairs(dict)
3-element Dictionaries.PairDictionary{String,Int64,Dictionary{String,Int64}}
"a" │ "a" => 1
"b" │ "b" => 2
"c" │ "c" => 3
julia> map(((k,v),) -> k^v, pairs(dict))
3-element Dictionary{String,String}
"a" │ "a"
"b" │ "bb"
"c" │ "ccc"You can sort a dictionary by value using the sort function (with the usual keyword arguments).
julia> sort(dict; rev = true)
3-element Dictionary{String, Int64}
"c" │ 3
"b" │ 2
"a" │ 1The sortkeys and sortpairs functions allows you to sort by a dictionary's key or key-value pair, respectively, and you can do in-place sorting on supported types with sort!, sortkeys! and sortpairs! (although these may be slower than the out-of-place algorithms).
The indices of a dictionary are unique, and form a set (in the mathematical sense). You can get the indices for any dictionary with the keys function.
julia> keys(dict)
3-element Indices{String}
"a"
"b"
"c"Whenever you call keys(::AbstractDictionary), you always receive an AbstractIndices in return.
Indices shares a similar implementation to Base.Set and can be used to perform set operations including union, intersect, setdiff, symdiff, and mutating counterparts. You can construct one from any iterable of unique elements.
julia> inds = Indices(["a", "b", "c"])
3-element Indices{String}
"a"
"b"
"c"You can also use the distinct function, which is similar to unique from Base, to construct indices where the input may not be unique.
julia> distinct([1,2,3,3])
3-element Indices{Int64}
1
2
3The distinct function may be considered as useful replacement of unique in many cases, as the unique function internally constructs a hashmap (Set) anyway before returning a Vector. However, a Indices iterates as fast as Vector and in many cases it can be useful to be able to map it into a dictionary.
Indices are insertable, so you can use insert! and delete! (or set! and unset!) to add and remove elements.
julia> insert!(inds, "d")
4-element Indices{String}
"a"
"b"
"c"
"d"
julia> delete!(inds, "d")
3-element Indices{String}
"a"
"b"
"c"One crucial property of AbstractIndices is that they are a subtype of AbstractDictionary (similar to how the keys of an AbstractArray are always AbstractArrays). But how can a set, or indices, be a dictionary? Under getindex, they form a map from each element to itself.
julia> inds["b"]
"b"Thus, if you iterate an AbstractIndices you are guaranteed never to get the same value twice, and the collection is a set. All the usual set operations like union, intersect, setdiff and symdiff are defined, as well as a newly exported predicate function disjoint(set1, set2) which returns true if set1 and set2 do not intersect/overlap according to an elementwise isequal check, and false otherwise (note that Dictionaries.disjoint is deprecated in favour of Base.isdisjoint in Julia 1.5 onwards).
Since all dictionaries have keys, even indices must have keys - and in this case keys(inds::AbstractIndices) === inds.
While the above properties for AbstractIndices may seem a little unnecessary at first, they lead to a variety of useful behavior.
If you wish to perform an operation on each element of a set, you can simply map or broadcast some indices, and return a dictionary. These operations cannot return an AbstractIndices since the mapping may or may not be one-to-one, so the results may not be distinct, while map/broadcast must preserve the number of elements and the keys.
julia> map(uppercase, inds)
3-element Dictionary{String,String}
"a" │ "A"
"b" │ "B"
"c" │ "C"
julia> inds .* "at"
3-element Dictionary{String,String}
"a" │ "aat"
"b" │ "bat"
"c" │ "cat"You can filter indices.
julia> filter(in(["a", "b"]), inds)
2-element Indices{String}
"a"
"b"To find the subset of dictionary indices/keys that satisfy some constraint on the values, use the findall function.
julia> dict
3-element Dictionary{String,Int64}
"a" │ 1
"b" │ 2
"c" │ 3
julia> inds2 = findall(isodd, dict)
2-element Indices{String}
"a"
"c"And, finally, one useful thing you can do with indices is, well, indexing. Non-scalar indexing of dictionaries is a little more complicated than that of arrays, since there is an ambiguity on whether the indexer is a single index or a collection of indices (for arrays, the scalar indices are integers (or CartesianIndexes) so this ambiguity is less of a problem). The Indexing.jl provides the getindices function to return a container with the same indices as the indexer, and this is re-exported here.
julia> getindices(dict, inds2)
2-element Dictionary{String,Int64}
"a" │ 1
"c" │ 3It has been suggested to make the syntax dict.[inds2] available in Julia in the future for unambiguous non-scalar indexing.
Lazy non-scalar indexing may be achieved, as usual, with the view function.
julia> view(dict, inds2)
2-element DictionaryView{String,Int64,Indices{String},Dictionary{String,Int64}}
"a" │ 1
"c" │ 3Boolean or "logical" indexing is also ambiguous with scalar and non-scalar indexing. Luckily, the findall function is a convenient way to convert a Boolean-valued dictionary into indices, which we can use with getindices:
julia> isodd.(dict)
3-element Dictionary{String,Bool}
"a" │ true
"b" │ false
"c" │ true
julia> getindices(dict, findall(isodd.(dict)))
2-element Dictionary{String,Int64}
"a" │ 1
"c" │ 3(Who knows - maybe we need syntax for this, too?)
The UnorderedDictionary container is another hash-based dictionary, but unlike Dictionary the order of elements is not defined. Internally, it has a slightly optimized version of the implementation of the Dict built into Julia, but supports the AbstractDictionary interface as well as tokens. It can be a bit faster than Dictionary for workloads focussing on insertion and deletion - such as when building a cache where iteration order and speed are unimportant.
The ArrayDictionary container is a simple, iteration-based dictionary that may be faster for smaller collections. It's keys are the corresponding ArrayIndices type. By default these contain Vectors which support mutation, insertion and tokenization, but they can contain other arrays such as SVectors (which make for good statically-sized dictionaries, with similarities with Base.ImmutableDict).
There is a FillDictionary container which lazily maps every key to the same value (only keeping a single copy of the value).
Indices that are based on sort ordering instead of hashing (both in a dense sorted form and as a B-tree or similar) are also planned.
The similar function is used to create a dictionary with defined indices, but undefined values that can be set/mutated after the fact. similar(dict, T) creates a container with the same indices as dict and, optionally, a new element type.
julia> similar(dict, Vector{Int})
3-element Dictionary{String,Array{Int64,1}}
"a" │ #undef
"b" │ #undef
"c" │ #undefThe behaviour is the same if dict is an AbstractIndices - you always get a dictionary with settable/mutable elements. Preserving the indices using similar and setting the values provides a huge performance advantage compared to iteratively constructing a new dictionary via insertion (see the bottom of this README).
On the other hand, values can be initialized with the fill(value, dict) function.
julia> fill(42, dict)
3-element Dictionary{String,Int64}
"a" │ 42
"b" │ 42
"c" │ 42The fill function can optionally define a wider type than the value, helpful for if you want to assign a default value like missing but allow this to be updated later.
julia> fill(missing, dict, Union{Missing, Int64})
3-element Dictionary{String,Union{Missing, Int64}}
"a" │ missing
"b" │ missing
"c" │ missingFunctions zeros, ones, falses and trues are defined as a handy alternative to the above in common cases, as are rand and randn.
julia> zeros(dict)
3-element Dictionary{String,Float64}
"a" │ 0.0
"b" │ 0.0
"c" │ 0.0
julia> zeros(UInt8, dict)
3-element Dictionary{String,UInt8}
"a" │ 0x00
"b" │ 0x00
"c" │ 0x00Note that the indices of the output are not guaranteed to be mutable/insertable - in fact, in the current implementation inserting or deleting indices to the output of the above can corrupt the input container (Julia suffers similar restrictions with AbstractArrays with mutable indices, for example changing the size of the indices of a SubArray can lead to corruption and segfaults). This also holds true for the output of map, broadcast, getindices, similar, zeros, ones, falses and trues. If you want a new container with indices you can insert, by sure to copy the indices first, or use empty instead.
The empty function will create an insertable container which is "similar" to the input, with zero elements and the specified type for the indices and values.
empty(x, I)constructs an empty indices (whetherxis a dictionary or indices).empty(x, I, T)constructs an empty dictionary (whetherxis a dictionary or indices).empty(x)constructs an empty container - indices ifxare indices, and a dictionary ifxis a dictionary.
This section will be of primary interest to developers who wish to understand the internals to Dictionaries.jl or create their own custom dictionary types.
The common supertype to this package is AbstractDictionary{I, T}, which models an indexable container. To implement a simple AbstractDictionary all you need to implement is:
getindex(::AbstractDictionary{I, T}, ::I) --> Tkeys(::AbstractDictionary{I, T}) --> AbstractIndices{I}isassigned(::AbstractDictionary{I, T}, ::I) --> Bool
Indexable containers in Julia have keys, which form a "set" in the mathematic sense of a collection of distinct elements. The keys of an AbstractDictionary{I, T} must have type AbstractIndices{I}. These form a set because no two elements in an AbstractIndices can be isequal. To implement a simple index type, you need to provide:
- The
iterateprotocol, returning unique values of typeI. in, such thatin(i, indices)implies there is an element ofindiceswhichisequaltoi.- Either
length, or overrideIteratorSizetoSizeUnknown.
Indices themselves are also dictionaries (much like the indices of AbstractArrays are also AbstractArrays), and we have the subtyping relationship AbstractIndices{I} <: AbstractDictionary{I, I}. Indexing an AbstractIndices is always an identity operation, such that indices[i] === i. The keys function is also an identity operation on indices (keys(indices::AbstractIndices) === indices) and therefore idempotent on dictionaries (keys(keys(dict::AbstractDictionary)) === keys(dict)).
Indexing an AbstractDictionary follows the interface provided by the Indexing.jl package. Since the indices of a dictionary may be of arbitrary type (including being a container such as an array or a dictionary), a function distinct to getindex is required to indicate non-scalar indexing.
The expression dict3 = getindices(dict1, dict2) follows the following simple rules:
- The output indices match the indexer, such that
issetequal(keys(dict3), keys(dict2)). - The values of
dict3come directly fromdict1, such thatdict3[i] === dict1[dict2[i]]for alli in keys(dict2).
Non-scalar indexing is simplified such that it is essentially getindices(dict1, dict2) = map(i -> dict1[i], dict2). Note also that getindices(dict, keys(dict)) has the same keys and values as dict, and is synonymous with getindices(dict, :).
These rules match those for AbstractArray, including offset arrays. The view function will work similarly, and the setindices! function from Indexing.jl is already implemented (see mutation, below).
Many dictionary types support setting or mutating the the values of the elements. To support mutation, an AbstractDictionary should implement:
issettable(::AbstractDictionary)(returningtrue)setindex!(dict::AbstractDictionary{I, T}, ::T, ::I)(returningdict)
The issettable function is a trait function that indicate whether an AbstractDictionary supports setindex!.
Because the idempotency property of AbstractIndices, indices always have immutable values - but indices can be inserted or deleted (see below).
If arbitrary indices can be added to or removed from an AbstractDictionary, one needs to implement:
isinsertable(::AbstractDictionary)(returningtrue)insert!(dict::AbstractDictionary{I, T}, ::I, ::T)(returningdict)delete!(dict::AbstractDictionary{I, T}, ::I)(returningdict)
The insert! and delete! always create or remove indices. Calling insert! when an index already exists will throw an error, as will attempting to delete! an index that does not exist. The function set! is provided as an "upsert" (update or insert) operation. Similarly, unset! function can be used to ensure a given index does not exist. The get! function works as in Base.
NOTE: setindex! can never create new indices, unlike with Julia's AbstractDict (and many other programming languages!). Always use set! to perform an "upsert" operation. This change may seem inconvenient at first, but it is similar to AbstractArray and how Julia differs from MATLAB in requiring one to explicitly push! to the end of a vector (a much less bug-prone pattern).
AbstractIndices may also be insertable, by implementing:
isinsertable(indices)(returningtrue)insert!(indices, i)- add new indexitoindices(will error if index exists)delete!(indices, i)- remove an existing indexifromindices(will error if index does not exist).
The set! and unset! functions behave as expected, as do union!, intersect!, setdiff! and symdiff!. Since indices iterate values, the filter! function can programmatically trim back a set of indices.
To make operations on dictionaries fast, we need to avoid unnecessary lookups into the dictionary and operations like recomputations of hashes. The token interface makes many things more efficient, especially co-iteration of similar containers containing identical keys.
A token is a more efficient way of referring to an element of indices. Using tokens may
help avoid multiple index lookups for a single operation.
A tokenizable indices must implement:
istokenizable(indices)(returningtrue)tokentype(indices) --> T::Typeiteratetoken(indices, s...)iterates the tokens ofindices, likeiterategettoken(indices, i) --> (hasindex::Bool, token)gettokenvalue(indices, token)returning the value of the index attoken
An isinsertable tokenizable indices must implement
gettoken!(indices, i) --> (hadtoken::Bool, token)deletetoken!(indices, token) --> indices
An tokenizable dictionary must implement:
istokenizable(dict)(returningtrue)keys(dict)must beistokenizableand share tokens withdictgettokenvalue(dict, token)returning the dictionary value attokenistokenassigned(dict, token) --> Bool
An issettable tokenizable dictionary must implement:
settokenvalue!(dict, token, value)
An isinsertable tokenizable dictionary must implement:
gettoken!(dict, i) --> (hadtoken::Bool, token)deletetoken!(dict, token) --> dict
When two-or-more dictionaries share the same tokens, co-iterating through their matching
elements becomes much more efficient. By default, the similar function on Indices
and ArrayIndices does not make a copy of the indices. When performing an operation such as
map!(f, d_out, d_in), a check of keys(d_out) === keys(d_in) lets us know that the
tokens are equivalent with a constant-time operation. When this is the case, the map!
operation can skip lookup entirely, performing zero calls to hash and dealing with hash
collisions.
A quick benchmark verifies the result.
julia> using Dictionaries, BenchmarkTools
julia> d1 = Dictionary(1:10_000_000, 10_000_000:-1:1);
julia> d2 = d1 .+ 1;
julia> @btime map(+, d1, d2);
25.712 ms (20 allocations: 76.29 MiB)The copy below makes keys(d1) !== keys(d2), disabling token co-iteration. It still uses
an iterative approach rather than using multiple hash-table lookups per element, so it's
relatively snappy.
julia> @btime map(+, d1, $(copy(d2)));
61.615 ms (20 allocations: 76.29 MiB)For a comparative baseline benchmark, we can try the same with dense vectors.
julia> v1 = collect(10_000_000:-1:1);
julia> v2 = v1 .+ 1;
julia> @btime map(+, v1, v2);
27.587 ms (5 allocations: 76.29 MiB)Here, the vector results are in line with the dictionary co-iteration!
Using insertion, instead of preserving the existing indices, is comparatively slow.
julia> function f(d1, d2)
out = Dictionary{Int64, Int64}()
for i in keys(d1)
insert!(out, i, d1[i] + d2[i])
end
return out
end
f (generic function with 1 method)
julia> @btime f(d1, d2);
2.819 s (10000091 allocations: 668.42 MiB)Unfortunately, insertion appears to be the idiomatic way of doing things with Base.Dict.
Compare the above to:
julia> dict1 = Dict(pairs(d1)); dict2 = Dict(pairs(d2));
julia> function g(d1, d2)
out = Dict{Int64, Int64}()
for i in keys(d1)
out[i] = d1[i] + d2[i]
end
return out
end
g (generic function with 1 method)
julia> @btime g(dict1, dict2);
9.507 s (72 allocations: 541.17 MiB)The result is similar with generators, which is possibly the easiest way of dealing with
Base.Dict.
julia> @btime Dict(i => dict1[i] + dict2[i] for i in keys(dict1));
13.046 s (89996503 allocations: 2.02 GiB)This represents a 500x speedup between the first example with Dictionary to this last
example with Base.Dict.