# How to define \`:(==)\` on user-defined types for \`Set\` operations
**URL:** https://discourse.julialang.org/t/how-to-define-on-user-defined-types-for-set-operations/126391
**Category:** General Usage
**Tags:** sets
**Created:** [February 27, 2025, 4:27pm UTC](https://discourse.julialang.org/t/how-to-define-on-user-defined-types-for-set-operations/126391 "2025-02-27T16:27:08Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@drbergman](https://discourse.julialang.org/u/drbergman)
#### Post date: [February 27, 2025, 4:27pm UTC](https://discourse.julialang.org/t/how-to-define-on-user-defined-types-for-set-operations/126391/1 "2025-02-27T16:27:08Z")
I want to create a `Set{MyType}` object, but use a custom `Base.:(==)` to compare instances of `MyType`. It seems that the dispatch of `in` is not making use of `Base.:(==)` (or `Base.isequal` for that matter) _despite_ what the documentation says. Here is a MWE:
```julia
struct MyType
a::Int
b::Int
end
Base.:(==)(x::MyType, y::MyType) = x.a == y.a
x = MyType(1, 2)
y = MyType(1, 3)
@assert x == y
S = Set{MyType}()
push!(S, x)
@assert x in S
@assert y in collect(S)
@assert !(y in S) # this one is unexpected / I would want this to actually return true
push!(S, y)
@assert length(S) == 2 # a new element has been added
@assert y in S # y is now in S
```
I know there have been a couple threads on this before ([here](https://discourse.julialang.org/t/in-operator-for-set-of-sets/50474) and [here](https://discourse.julialang.org/t/create-a-user-type-set/75420)) but these seem to be centered around mutable elements in a `Set`.
My use-case is not so performance critical, so I’ll probably just use a `Vector{MyType}` and check if `y in S` before `push!`ing, but it would be helpful if the docs could be updated to reflect what seems (to the best of my testing/comprehension of the docs) unexpected behavior.
Finally, here’s a MWE also with overloading `Base.isequal`. The output makes it clear that only after collecting does `in` use `:(==)`.
```julia
struct MyType
a::Int
b::Int
end
function Base.:(==)(x::MyType, y::MyType)
println("USING UPDATED :(==)")
x.a == y.a
end
function Base.isequal(x::MyType, y::MyType)
println("USING UPDATE isequal")
x.a==y.a
end
x = MyType(1, 2)
y = MyType(1, 3)
@assert x == y
S = Set{MyType}()
push!(S, x)
@assert x in S
@assert y in collect(S)
@assert !(y in S) # this one is unexpected / I would want this to actually return true
push!(S, y)
@assert length(S) == 2 # a new element has been added
@assert y in S # y is now in S
```
#### `versioninfo()`
```julia-auto
Julia Version 1.11.3
Commit d63adeda50d (2025-01-21 19:42 UTC)
Build Info:
Official https://julialang.org/ release
Platform Info:
OS: macOS (arm64-apple-darwin24.0.0)
CPU: 16 × Apple M4 Max
WORD_SIZE: 64
LLVM: libLLVM-16.0.6 (ORCJIT, apple-m1)
Threads: 1 default, 0 interactive, 1 GC (on 12 virtual cores)
Environment:
JULIA_EDITOR = code
JULIA_NUM_THREADS =
```
---
### Author:  [@mbauman](https://discourse.julialang.org/u/mbauman)
#### Post date: [February 27, 2025, 4:30pm UTC](https://discourse.julialang.org/t/how-to-define-on-user-defined-types-for-set-operations/126391/2 "2025-02-27T16:30:28Z")
That second topic — [Create a user-type set](https://discourse.julialang.org/t/create-a-user-type-set/75420) — is exactly your situation. The mutability is irrelevant here. When you re-define [`==`](https://docs.julialang.org/en/v1/base/math/#Base.:==), you must also symmetrically redefine [`hash`](https://docs.julialang.org/en/v1/base/base/#Base.hash).
---
### Author:  [@drbergman](https://discourse.julialang.org/u/drbergman)
#### Post date: [March 3, 2025, 3:51pm UTC](https://discourse.julialang.org/t/how-to-define-on-user-defined-types-for-set-operations/126391/3 "2025-03-03T15:51:34Z")
Thank you so much @mbauman! I guess what I really needed was confirmation that this was the right approach. Looking at that post again, I decided to use `AutoHashEquals.jl` and defined
```julia
using AutoHashEquals
@auto_hash_equals fields = (a,) struct MyType
a::Int
b::Int
end
```
And it’s working perfectly!