Splitting up Numerical Text Data

I am using the webread function to obtain numerical text data from a webpage. Each row of data represents a different weather element while each column represents the forecast time. I have had success spliting up the numerical data from rows like TMP, DPT, and WSP (Temperature, Dew point, and Wind Speed) where the numbers usually only have two digits. However, when rows like TMP, SKY, VIS (Temperature, Sky Cover, Visiblity) have three digits, the text may appear: '89 93 96 9910010010098 96 94 90' and using str2num will output 89 93 96 9910010010098 96 94 90 when I want it to read 89 93 96 99 100 100 100 98 96 94 90 so I can plot the values. I've been using str2num for two digit numbers and it works great. Given that the string of numerical text can change for a different forecast time period and location, how can I write a section of code that splits up this data?

[Cris LaPierre]

The issue here appears to be there is no clear delimiter between your numbers. Even though this data appears to be fixed width, the format changes. With it not being consistant, there is no simple fix I'm aware of.

Are you positive you copied the data correctly? Is there any chance it should be this?

'89 93 96 99100100100 98 96 94 90'

[Mathieu NOE]

hello

can you share some text data with your different data length

tx

[Dalton Van Stratten]

I provided a screen capture below to help you become familiar with the format.

The problem rows that I am trying to utilize are TMP when temperature is equal to or exceeds 100 degrees, SKY when sky cover is 100%, PZR, PSN, PPL, PRA when probablity of precipitation falling as a specific type is 100%, and visibility when visibility is 100. Unfortunately, the format is basically broken whenever there is a 100 in the data.

And yes, the format you mentioned would be the correct example. It should be '89 93 96 99100100100 98 96 94 90' as an example. Does that change things?

The delimiter is a space in this case, but when there is a 100 in the data, there is no delimiter. Let's say I want to split the text '99100'. One could split that text at the 9 and the 1 but if we had '98100' that would have to be split at the 8 and the 1. Even if that is possible for any number combination, the next string could be '100100' which would be split at the 0 and the 1.

The only way I can think of solving this problem is using something like IF str2num('string') <= 100 check to see where 'string' should be logically split.

[Cris LaPierre]

Yes, that does change things. It means the data is fixed width, which makes it possible to separate. I therefore believe it is actually ' 89 93 96 99100100100 98 96 94 90'.

Answer below.

Answer 1

[Cris LaPierre]

t = ' 89 93 96 99100100100 98 96 94 90';
T = textscan(t,'%3s','Whitespace','');
T = str2double(T{:})

T = 11×1

89 93 96 99 100 100 100 98 96 94

●

[J. Alex Lee]

is that any better than

T = cell2mat(textscan(t,'%3f','Whitespace',''));

?

[Cris LaPierre]

I had tried %f, but for me, it wasn't capturing the numbers correctly.

t = ' 89 93 96 99100100100 98 96 94 90';
T = cell2mat(textscan(t,'%3f','Whitespace',''))

T = 11×1

89 93 96 991 1 1 0 98 96 94

●