How to replace string values in a column of a table into numeric?

Question

Vaishali Ravi 2020-8-3

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/574351-how-to-replace-string-values-in-a-column-of-a-table-into-numeric

编辑： Hamidreza 2022-9-21

I have a table X

It has a column named 'fruit' with string values such as {'apple','orange','grapes'}, There are 13 different string values in the column 'fruit'

Now I need to change it as if apple or orange then 1, grapes then 2 and so on.

How can I do that?

4 个评论
显示 2更早的评论隐藏 2更早的评论

Walter Roberson 2020-8-3

you need to replace the entire table variable if you are changing datatype.

It is often easier to add a new variable with the derived data and either leave the old data or delete the variable.

By the way, use ismember and the second output to find the index of the string. Use the index of the string to index into a vector that maps to numeric values, the thus allowing you to map orange and apple to the same number.

This can be done in a small number of vectorized lines m

Vaishali Ravi 2020-8-3

can you giva a sample code on how to use ismember?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2020-8-3

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/574351-how-to-replace-string-values-in-a-column-of-a-table-into-numeric#answer_474730

fruit_types = {'apple', 'orange', 'grape', 'banana', 'kumquat', 'peach', 'tomato', 'strawberry', 'pineapple', 'kiwi', 'blueberry', 'durian', 'avacodo'};
fruit_value = [      1,        1,       2,      -4,        5.3,      1         3          5.3,           11,      2,        -5.3,        1,         4];
[wasfound, idx] = ismember(X.fruit, fruit_types);
f_values = nan(length(idx), 1));
f_values(wasfound) = fruit_value(idx(wasfound));
X.fruit_value = f_values;

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Hamidreza 2022-9-21

编辑：Hamidreza 2022-9-21

When I run your code I get the error message below:

Error using cell/ismember (line 34)

Input A of class double and input B of class cell must be cell arrays of character vectors, unless one is a character vector.

Any advice to resolve it?

Thanks,

Hamid

请先登录，再进行评论。

Answer 2

Cris LaPierre 2020-8-3

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/574351-how-to-replace-string-values-in-a-column-of-a-table-into-numeric#answer_474478

I would first convert fruit to a categorical, then I would rename the categories.

X.fruit = categorical(X.fruit);
X.fruit = renamecats(X.fruit,{'apple','orange','grapes'},{'1','2','3'});

4 个评论
显示 2更早的评论隐藏 2更早的评论

Vaishali Ravi 2020-8-4

Actually I tried str2double, it produced Nan values.

Cris LaPierre 2020-8-4

You have to combine all 3 lines of code:

Convert to categorical
Rename categories
Convert categories to double

% set up a table variable with 20 repeating fruit names
fruit = {'apple','orange','grapes'}';
ind = randi(3,20,1);
fruit = fruit(ind);
X = table(fruit)

Here you see the text values for fruit

X =
  20×1 table
      fruit   
    __________
    {'apple' }
    {'apple' }
    {'grapes'}
    {'grapes'}
    {'orange'}
    {'grapes'}
    {'grapes'}
    {'apple' }
    {'orange'}
    {'orange'}
    {'grapes'}
    {'apple' }
    {'orange'}
    {'orange'}
    {'orange'}
    {'grapes'}
    {'apple' }
    {'grapes'}
    {'orange'}
    {'apple' }

Now convert the text to numbers

% Change names to numbers
X.fruit = categorical(X.fruit);
X.fruit = renamecats(X.fruit,{'apple','orange','grapes'},{'1','2','3'});
X.fruit = str2double(string(X.fruit))

Now you see the fruit values are numbers

X =
  20×1 table
    fruit
    _____





















Use summary to confirm fruit is now a double.

>> summary(X)
Variables:
    fruit: 20×1 double
        Values:
            Min          1   
            Median       2   
            Max          3 