# Dividing matrix into sections
**URL:** https://discourse.julialang.org/t/dividing-matrix-into-sections/10428
**Category:** Performance
**Created:** [April 19, 2018, 7:10am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428 "2018-04-19T07:10:38Z")
**Posts on this page:** 13
**Page:** 1
### Author:  [@abshej](https://discourse.julialang.org/u/abshej)
#### Post date: [April 19, 2018, 7:10am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/1 "2018-04-19T07:10:38Z")
How to divide a matrix like _m_ x _n_ or an Array like _m_ x _n_ x 1 into topographic sections? Imagine dividing a square into smaller squares.
And operating individually on those sections.
Edit: **The actual problem is that I want to know the indices of each element of each section the way they are stored in the parent matrix**
---
### Author:  [@Tamas\_Papp](https://discourse.julialang.org/u/Tamas_Papp)
#### Post date: [April 19, 2018, 7:31am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/2 "2018-04-19T07:31:52Z")
Maybe I am missing something, but possibly with indexing and views? Eg
```julia
julia> A = zeros(3, 3);
julia> Av = @view A[1:2, 1:2];
julia> Av .= ones(2, 2);
julia> A
3×3 Array{Float64,2}:
1.0 1.0 0.0
1.0 1.0 0.0
0.0 0.0 0.0
```
---
### Author:  [@abshej](https://discourse.julialang.org/u/abshej)
#### Post date: [April 19, 2018, 7:40am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/3 "2018-04-19T07:40:20Z")
But how will we do this for the entire matrix? Say you wish to divide it into 6 squares.
---
### Author:  [@Tamas\_Papp](https://discourse.julialang.org/u/Tamas_Papp)
#### Post date: [April 19, 2018, 8:06am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/4 "2018-04-19T08:06:08Z")
You specify the division and map it to views. (If you have a concrete example of what you want, I could help with code).
---
### Author:  [@abshej](https://discourse.julialang.org/u/abshej)
#### Post date: [April 19, 2018, 8:43am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/5 "2018-04-19T08:43:31Z")
```julia
a = rand(1:100, 25, 25, 1)
i, j, one = size(a)
for m in 1:i, for n in 1:j
display(a[max(1, m-4):min(25, m+4), max(1, n-4):min(25, n+4)])
end
```
---
### Author:  [@Tamas\_Papp](https://discourse.julialang.org/u/Tamas_Papp)
#### Post date: [April 19, 2018, 8:55am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/6 "2018-04-19T08:55:09Z")
```julia
a = rand(1:100, 25, 25, 1)
i, j, _ = size(a) # replace `one`, don't overwrite that function
V = [@view(a[max(1, m-4):min(25, m+4), max(1, n-4):min(25, n+4), :]) # added a :
for m in 1:i for n in 1:j] # note nesting order, is this what you want?
```
This gives you a vector of views, you can modify the comprehension to make a matrix if that’s what you need.
---
### Author:  [@abshej](https://discourse.julialang.org/u/abshej)
#### Post date: [April 19, 2018, 9:12am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/7 "2018-04-19T09:12:11Z")
That’s not what I want. It’s just a workaround to store views which I was aware of. I was looking for a function to divide a matrix into mostly equal squares so that I can run a function over those individual sections.
**But now I realize that the actual problem is that I still want to somehow know the indices of each element of each section the way they are stored in the parent matrix**
---
### Author:  [@Tamas\_Papp](https://discourse.julialang.org/u/Tamas_Papp)
#### Post date: [April 19, 2018, 9:17am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/8 "2018-04-19T09:17:04Z")
Well, there is `Base.parent` and `Base.parentindexes`. I don’t know the details of your calculation, but this may be a very roundabout way of going about it.
---
### Author:  [@Evizero](https://discourse.julialang.org/u/Evizero)
#### Post date: [April 19, 2018, 9:39am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/9 "2018-04-19T09:39:14Z")
Take a look at [GitHub - JuliaArrays/TiledIteration.jl: Julia package to facilitate writing mulithreaded, multidimensional, cache-efficient code](https://github.com/JuliaArrays/TiledIteration.jl)
---
### Author:  [@abshej](https://discourse.julialang.org/u/abshej)
#### Post date: [April 19, 2018, 9:43am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/10 "2018-04-19T09:43:03Z")
Thanks a lot.
---
### Author:  [@mkborregaard](https://discourse.julialang.org/u/mkborregaard)
#### Post date: [April 20, 2018, 8:34am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/11 "2018-04-20T08:34:45Z")
Or alternatively [https://github.com/JuliaImages/ImageFiltering.jl](https://github.com/JuliaImages/ImageFiltering.jl) which are general array operations, not just for images.
---
### Author:  [@Seif\_Shebl](https://discourse.julialang.org/u/Seif_Shebl)
#### Post date: [April 20, 2018, 11:25pm UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/12 "2018-04-20T23:25:30Z")
For square matrices of size n\times n, for example, this can be easy:
```julia
n = 1024
bSize = 16
A = rand(n,n)
for j = 1:bSize:n, i = 1:bSize:n
tempArray = @view A[i:i+bSize-1,j:j+bSize-1]
# do calculations on tempArray
end
```
For general rectangular matrices m\times n, you need a `row_bSize` and another `col_bSize`, such that `row_bSize` divides m and `col_bSize` divides n but the idea is the same. If the blocks need to be squares, then `row_bSize = col_bSize` and m or n are divisible my either of them.
---
### Author:  [@dlfivefifty](https://discourse.julialang.org/u/dlfivefifty)
#### Post date: [April 21, 2018, 6:01am UTC](https://discourse.julialang.org/t/dividing-matrix-into-sections/10428/13 "2018-04-21T06:01:55Z")
Or [https://github.com/JuliaArrays/BlockArrays.jl](https://github.com/JuliaArrays/BlockArrays.jl)