# How to assign subsets of a large dataset into different arrays within a loop
**URL:** https://discourse.julialang.org/t/how-to-assign-subsets-of-a-large-dataset-into-different-arrays-within-a-loop/5218
**Category:** Data
**Tags:** juliacomputing
**Created:** [August 4, 2017, 7:55am UTC](https://discourse.julialang.org/t/how-to-assign-subsets-of-a-large-dataset-into-different-arrays-within-a-loop/5218 "2017-08-04T07:55:41Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@Devi\_Sree](https://discourse.julialang.org/u/Devi_Sree)
#### Post date: [August 4, 2017, 7:55am UTC](https://discourse.julialang.org/t/how-to-assign-subsets-of-a-large-dataset-into-different-arrays-within-a-loop/5218/1 "2017-08-04T07:55:41Z")
I am using for loop that divides a large array say `arr` and assign the subsets into arrays ` arr1, arr2.`.
I am stuck at assigning subsets to arrays within a loop.
My whole datastet `arr` contains 5 rows and 5 columns.
```
arr= [1 3 4 5 6;2 5 6 7 8;2 3 1 4 6 ;8 6 5 3 2;1 5 8 5 4]
```
I need to this split into `arr1, arr2, arr3` as
```
arr1=[1 3 4 5 6;2 5 6 7 8]
arr2=[2 3 1 4 6 ;8 6 5 3 2]
arr3=[1 5 8 5 4] // remaining
```
I want to do this splitting inside loop so that the three lines get reduced to one line.
---
### Author:  [@Nosferican](https://discourse.julialang.org/u/Nosferican)
#### Post date: [August 4, 2017, 8:27am UTC](https://discourse.julialang.org/t/how-to-assign-subsets-of-a-large-dataset-into-different-arrays-within-a-loop/5218/2 "2017-08-04T08:27:47Z")
If you could provide a more concrete example it would be helpful.
As an example I will assume the array is a matrix.
```julia
srand(0)
using Distributions
arr = rand(100, 5)
index = Distributions.sample(1:5, 100)
output = map(group -> arr[group .== index,:], unique(index))
```
In this example `index` is just a vector which indicates to which group the row (splitting by the first dimension) belongs to. You could use a _for loop_, but for most instances `map` will be sufficiently efficient. I prefer `map` to list comprehension since it will preserve the type (`Matrix`). If you just need to perform a computation on each subset, you could calculate it on the fly or store the `Vector` of `Matrix` and perform the calculations afterwards.