Find matched string in table

Hi,

To find data matching certain conditions in a table we use:

rows = (T.Smoker==true & T.Age<40);

What if the T.Smoker field was not a logical but a string? 'yes' or 'no'.

rows = (T.Smoker=='yes' & T.Age<40);

This later code does not work. How could I make it work so the condition matches a certain string?

Thank you,

TD

Answer 1

[Steven Lord]

Compare the text data stored in the table with a string array. Let's build a sample using the example from the documentation for the table function.

load patients
patients = table(LastName,Gender,Age,Height,Weight,Smoker,Systolic,Diastolic);

Let's extract patients under 40, separated by gender.

malePatientsUnder40 = patients(patients.Gender == "Male" & patients.Age < 40, :)
femalePatientsUnder40 = patients(patients.Gender == "Female" & patients.Age < 40, :)

Let's check against the full list of patients under 40.

patientsUnder40 = patients(patients.Age < 40, :);

Combine the male and female patients into one larger table and compare it against the full list.

isequal(sortrows([malePatientsUnder40; femalePatientsUnder40]), sortrows(patientsUnder40))

[The Finnish Rein Deer]

Thanks. It's funny, when using only

patients(patients.Gender == "Female")

It returns error "Subscripting a table using linear indexing (one subscript) or multidimensional indexing (three or more subscripts) is not supported. Use a

row subscript and a variable subscript.".

[Image Analyst]

Yeah, that's because you forgot to include the comma and colon to indicate "all columns", as Steve showed in his excellent answer. I wonder if this would work if the column were character arrays instead of strings, meaning 'Female' instead of "Female" (single quotes instead of double quotes).

Since you ended up using his answer instead of mine, you can unaccept mine and accept his Answer instead.

[Steven Lord]

If you build the patients table using the patients.mat file the Gender variable is actually a cellstr, a cell array containing char vectors. It's not a string array, though you could use the cellstr to make one.

Comparing char arrays with string or categorical arrays we try to do "the nice thing" and and treat the text more like "words" rather than "bunches of characters". See this documentation page for a discussion about comparing char and string arrays and this one for char and categorical.

How would these work? Let's rebuild the patients table.

load patients
patients = table(LastName,Gender,Age,Height,Weight,Smoker,Systolic,Diastolic);

What type is the Gender variable in patients? It's a cellstr.

class(patients.Gender)
iscellstr(patients.Gender)

Since in this data set Gender only takes two values, we can store and display them as two categories using a categorical array. We can match a categorical array using a char vector, a string, or a categorical value. [I'm switching the Age threshold to 30 just so the filtered table arrays are shorter.]

patients.GenderCat = categorical(patients.Gender);
F1 = patients(patients.GenderCat == 'Female' & patients.Age < 30, :)
F2 = patients(patients.GenderCat == "Female" & patients.Age < 30, :)

FCategory = patients.GenderCat(3) % The first Female patient is in row 3
F3 = patients(patients.GenderCat == FCategory & patients.Age < 30, :)

Or you can turn that cellstr data into string.

patients.GenderStr = string(patients.Gender);
F4 = patients(patients.GenderStr == "Female" & patients.Age < 30, :)
F5 = patients(patients.GenderStr == 'Female' & patients.Age < 30, :)

[The Finnish Rein Deer]

Thank you so much! All clear.

Answer 2

[Fangjun Jiang]

isequal(T.Smoker, 'yes')
strcmpi(T.Smoker, 'yes')

[The Finnish Rein Deer]

Thanks. Isequal does not work, and strcmpi is better to replace for strcmp, since the former would match independently of case.

Answer 3

[Image Analyst]

Try contains:

rows =  contains(T.Smoker, 'yes') & T.Age<40;

[The Finnish Rein Deer]

Actually, I realized that the code I suggested would work if only I had changed simple quotes for double ones!

rows = (T.Smoker=='yes' & T.Age<40); % does not work
rows = (T.Smoker=="yes" & T.Age<40); % does work