# How to take slice of array without it turning into vector

**URL:** <https://discourse.julialang.org/t/how-to-take-slice-of-array-without-it-turning-into-vector/36352>\
**Category:** New to Julia\
**Created:** [March 22, 2020, 5:22pm UTC](https://discourse.julialang.org/t/how-to-take-slice-of-array-without-it-turning-into-vector/36352 "2020-03-22T17:22:15Z")\
**Posts on this page:** 2\
**Page:** 2

<div class="post-metadata">

**Author:** ![DNF](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/dnf/32/10191_2.png) [@DNF](https://discourse.julialang.org/u/DNF)\
**Post date:** [March 27, 2020, 2:26pm UTC](https://discourse.julialang.org/t/how-to-take-slice-of-array-without-it-turning-into-vector/36352/21 "2020-03-27T14:26:23Z")

</div>

> [@kw\_martin](#):
>
> Gunnar, my understanding is it’s better to use columns for the variable that iterates in the inside loop. For filters, I tend to reserve this for containers that have elements changing with time. Hence, the use of rows for each stage of a cascade filter.

This isn’t really clear to me. Isn’t the ‘inner loop’ necessarily along the filter coefficients, not _across_ the different filters?

> [@kw\_martin](#):
>
> My personable opinion (not shared by others as far as I can tell), is that complex signal processing is actually cleaner and simpler than “real” signal processing.

Is this in reference to something in this thread?

---

<div class="post-metadata">

**Author:** ![kw\_martin](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/kw_martin/32/48868_2.png) [@kw\_martin](https://discourse.julialang.org/u/kw_martin)\
**Post date:** [March 27, 2020, 3:58pm UTC](https://discourse.julialang.org/t/how-to-take-slice-of-array-without-it-turning-into-vector/36352/22 "2020-03-27T15:58:36Z")

</div>

Re your second comment; not really relevant, just some  
background. Re your first comment: there is some choice here. For  
complex cascade filter banks, there are actually three loops; one  
for each filter in the bank (currently 256); one for each stage in  
each cascade filter, and one for each input sample that goes to  
all of the filter banks. The order in which each loop is done is  
one of the design choices; currently, I am focusing on just  
getting the filters working (yesterday I was successful in this  
regard), and I will leave optimization to later once I have all of  
the pieces in place for ECG analysis (just on clean signals for  
now).

[Previous page](https://discourse.julialang.org/t/how-to-take-slice-of-array-without-it-turning-into-vector/36352.md?page=1)