One measure greater than another?

Hi All,

I have a 697 x 38 which has different column headings such as 'startf, maxf, etc'.

I'm wanting to check if my measurements are true (an answer of 1 indiciting not true?) such that;

startf>fmax

startf<fmax etc..

Currently, I've sectioned off the columns of interests:

T = PWALLVOCSPARTSOFBIPHON
Wanted = T(:,'SFStartFreqHzoffundamental')
Wanted1 = T(:,'MaxFHzfundamental'

I now want to check if wanted is greater than wanted1

I tried:

Wanted > Wanted1 ans =

but thats just producing an error.

In addition, if the answers were shown to be untrue. Such as I have data of a start frequency that is lower than minimum frequency, how do i find which rows these will be located in?

[Walter Roberson]

Wanted = T{:,'SFStartFreqHzoffundamental'};
Wanted1 = T{:,'MaxFHzfundamental'};
mask = Wanted > Wanted1;

[Rachael Courts]

Thank-you for responding Walter!

Ran your code. Unfortunatley, I'm still receiving this error:

Undefined operator '>' for input arguments of type 'table'.

The aim is to retrieve the response of either 1 or 0 so I know if the columns (in comparison) are true or not

[Walter Roberson]

After you have extracted those variables from the table using {} indexing (important!) then what is class(Wanted) and class(Wanted1) ?

WIth your current code using () then you would be taking a subset of a table and the result would be a table, but {} indexing extracts the contents of a varialbe, and the only way it could be a table as a result is if you have tables stored inside of tables.

[Rachael Courts]

THANK YOU SO MUCH! parentheses are definitely important, thank-you for explaining it too.

Okay, so the code works perfectly!

The next step, to retrieve the answer of 0 or 1? then locating within the column what row has the incorrect value?

Any suggestions? :)

[Walter Roberson]

find()?

Answer 1

[Guillaume]

Note than an even simpler way to extract data from a table is with dot indexing:

Wanted = T.SFStartFreqHzoffundamental;  %equivalent to T{:,'SFStartFreqHzoffundamental'}
Wanted1 = T.MaxFHzfundamental;          %equivalent to T{:,'MaxFHzfundamental'}
mask = Wanted > Wanted1;

To keep only the portion of the table for whick mask is true:

Tinvalid = T(mask, :)  %Not that this returns a table since we're using () indexing

which can be done in only one line:

Tinvalid = T(T.SFStartFreqHzoffundamental > T.MaxFHzfundamental, :)

If you want to know the row indices, use find:

invalidrows = find(T.SFStartFreqHzoffundamental > T.MaxFHzfundamental)

[Rachael Courts]

Hi Guillaume,

Thank-you for answering! so when I run your code I receive this output:

Tinvalid =

0×38 empty table

invalidrows =

0×1 empty double column vector

I'm not sure how this shows whether it is a true (1) or false (0), is it the '0' in the '0 x1 empty double column vector' which indicates that all of the outputs are false? (which would be correct as this is a dummy run)

[Guillaume]

Well, yes you've got no invalid rows, so you get empty results. If you indeed had invalid rows, you'd see them.

The number of invalid rows is:

nnz(mask)

[Rachael Courts]

Thank you so much!