Sort Descend Scientific Notation Error

Hi,

I've got a vector of numbers. The vector has some numbers stored with and without scientific notation.

Eg 0.01 and 9.8e-04.

When i try sorting these numbers in descending order, the sort function puts 9.8e-05 above 0.01, even though 0.01 is larger.

Can anyone help me so it sorts from largest to smallest based on the expanded values? I've been trying to convert the numbers in the matrix from scientific notation to standard form, but changing format has not helped, and i haven't been able to get rid of the scientific notation properly.

I've looked through past questions / answers and have not found an answer that has worked for me.

Any advice is appreciated.

I believe the issue may be that the numbers are considered strings, but unsure how to convert string scientific notation to numbers.

Example data and code:

u =

[ APH 0.01

GPT 9.8e-04

CPH 6.4e-05]

[~,idx] = sort(u(:,2),'descend');

sortedu = u(idx,:);

sortedu looks like:

sortedu =

[GPT 9.8e-04

CPH 6.4e-05

APH 0.01]

Thanks,

Kiran

edit: updated code.

[Stephen Cobeldick]

Not much of your code makes sense. The row vector u is defined like this (so far no problems):

u  = [ 0.01 9.8e-04 6.4e-05 ];

but then you use indexing to take only the second column, which is a scalar:

>> u(:,2)
ans =  0.00098000

You then sort that scalar, where obviously the sort index of a scalar is one (and it makes absolutely no difference using 'ascend' or 'descend'), and this is easy to confirm:

>> [~,idx] = sort(u(:,2))
idx =  1

You then use the index 1 to select the first row of u. I guess you thought that your code did some sorting and that idx would change the order. In fact your code is simply equivalent to:

sortedu = u(1,:)

which because u is a row vector anyway is simply

sortedu = u

This is not the output that you showed us, which means that you did not actually give us the data and/or code that you used. If you do not show us what you are actually doing, then you make it hard for us to help you.

My suspicion is that you actually have a string array, not a numeric array as you showed us.

[Kiran Eswaran]

Apologies, you are correct. I tried simplifying the code/data and in doing so made it incorrect.

The initial row vector i defined (u) is actually the 2nd column of an array u (the first column is a list of stock tickers).

for example, u looks like:

u =

[ APH 0.01

GPT 9.8e-04

CPH 6.4e-05]

[~,idx] = sort(u(:,2),'descend');

sortedu = u(idx,:);

sortedu =

[GPT 9.8e-04

CPH 6.4e-05

APH 0.01]

Apologies for the initial confusion.

[Kiran Eswaran]

Adding on to that point, the initial stock data were stored in a cell. I extracted from the cell as a string (string(cellname)), and then converted the string using str2double, as str2num didn't work for me.

• str2double(string(cellname));

I guess a better way to frame my question:

How can i change the string scientific notation (eg 9.8e-04) to a number.

Cheers,

Kiran

[Stephen Cobeldick]

@Kiran Eswaran : please upload your original data in a .mat file, by clicking the paperclip button. Note that I wrote "original data": this means your original table u (or whatever it is called), before you tried to do any conversions.

Answer 1

[Jos (10584)]

You misuse the sort command. Simply:

sortedu = sort(u ,'descend')

would do :-)

Moreover, if you remove the semicolons, you would see that you have some strange results in between...

Ask yourself what u(:,2) does, and u(idx,:).

[Kiran Eswaran]

Apologies, i framed the initial code incorrectly and have updated it.

I think the issue is that the numbers are stored as strings in scientific notation form (eg 9.8e-04) and so i'm unsure how to convert this string scientific notation to a number. I believe solving that conversion will solve the sort issue.

Cheers,

Kiran