Find Best Array size for n elements

[ATE-EGNE]

I have a 1D array of N elements that I'd like to convert into a 2D array of elements nxm where n and m are within 1 of each other so that

N=1 -> n = 1 m = 1

N=2 -> n = 2 m = 1

N=(3-4) -> n = 2 m = 2

N=(5-6) -> n = 3 m = 2

N=(7-9) -> n = 3 m = 3

N=(10-12) -> n = 4 m = 3

ect.

There's definitely a formula to describe this, but I can't put it together for the life of me. Any help on the best way to do this in LabVIEW would be appreciated.

[Kevin_Price]

Only considered this briefly but how about:

Given N:

1. Take sqrt(N).  Round up to next higher integer, call result J.

2. Candidate A: J x (J-1) array if J*(J-1) >= N, else...

3. Candidate B: J x J array will always be big enough

-Kevin P

[Mancho00]

EDIT:  OOPS.  remove the divide by two

2nd Edit:  This won't guarantee n and m are within 1 of each other for higher values of N

[Minions]

If n can be equal to or greater than m by 1, then here would be my solution.

[kashiruvana]

What you describe actually can't be done; what you're really describing is getting an optimized array that is the smallest almost-square. You're still padding your data array because N cannot always (even in the examples you give) fill a rectangular array that is "close" to being square. I think the closest you're going to get is this:

[Minions]

Kashiruvana,

It can be done.  Once the m & n dimensions have been determined, initialize the m*n array with NaN then fill from the single dimension array.

[kashiruvana]

You're still padding the array, as I said in my response. You have to have a rectangular array, and even from the original examples, if N=10 and you want to put it into a 3x4 array, you're changing your original N=10 array into an N=12 array padded with two NaNs at the end. You don't have to stop and take that step, obviously; you can write your N=10 array into a 3x4 array initialized with NaNs and it happens along the way. But you can't take any 1D array of length N and put it into a 2D array of nxm without padding for any arbitrary N. For any example, you're padding with (nm)-N NaNs (or zeroes or whatever you pick).

[altenbach]

Reshaping pads the array with zeroes as needed.

Here's what I would do.

[kashiruvana]

Ooh, I like that. That's a good solution. (If I could mark it, I would!)

When I wrote mine, I was under a time crunch, so I was looking for something that would give me the mantissa and exponent in base 10 instead of base 2, and I didn't see anything like that. So I came up with what I have, which works but definitely makes me uncomfortable--I mean, what? Convert to a string, then pick a magic index to check the char-to-num value of that, then... If I came across that in someone else's code that I had to refactor or use, I'd be mad, lol. I like yours a lot better.

[RavensFan]

---

@kashiruvana wrote:

Ooh, I like that. That's a good solution. (If I could mark it, I would!)

---

Why can't you?  Just use the Accept as Solution button beneath the post.