Matlab中实现特定文本替换为数值的最优方法咨询

Hey there! Let's figure out the best way to handle this text-to-numeric replacement in MATLAB. Depending on what kind of data structure you're working with, here are a few optimal approaches that'll get the job done quickly:

Method 1: Using a Mapping Container (Clean for Cell/String Arrays) #

If your data is stored in a cell array or string array, using a mapping structure keeps your code organized, especially if you might add more categories later.

For cell arrays:

% Sample data
data = {'Mostly false', 'Mostly true', 'Definitely true', 'Mostly false'};

% Create a key-value map
value_map = containers.Map({'Mostly false', 'Mostly true', 'Definitely true'}, {1, 2, 3});

% Convert to numeric array
numeric_data = cellfun(@(x) value_map(x), data);

For modern string arrays (MATLAB R2016b+), use dictionary instead—it's more intuitive:

data = ["Mostly false", "Mostly true", "Definitely true", "Mostly false"];
value_map = dictionary(["Mostly false", "Mostly true", "Definitely true"], [1,2,3]);
numeric_data = value_map(data);

Method 2: Categorical Conversion (Great for Retaining Context) #

If you want to keep a link between the original text and numeric values (say, for later analysis or plotting), converting to a categorical variable is perfect. It lets you easily switch between text and numbers whenever needed.

% Convert to categorical with explicit value mapping
cat_data = categorical(data, ...
    {'Mostly false', 'Mostly true', 'Definitely true'}, ...
    [1, 2, 3]);

% Convert to numeric when needed
numeric_data = double(cat_data);

% You can still get back the original text with char(cat_data) or string(cat_data)

Method 3: Logical Indexing (Simple & Fast) #

For straightforward replacements with just a few categories, logical indexing is hard to beat—it's lightweight and easy to read.

For cell arrays:

numeric_data = zeros(size(data)); % Initialize a numeric array of the same size
numeric_data(strcmp(data, 'Mostly false')) = 1;
numeric_data(strcmp(data, 'Mostly true')) = 2;
numeric_data(strcmp(data, 'Definitely true')) = 3;

For string arrays, you can use direct equality checks instead of strcmp:

numeric_data = zeros(size(data));
numeric_data(data == "Mostly false") = 1;
numeric_data(data == "Mostly true") = 2;
numeric_data(data == "Definitely true") = 3;

Bonus: Handling Table Data #

If your data is in a MATLAB table (super common for datasets), you can adapt the categorical method to add a new numeric column:

% Assume your table has a column named 'Response'
table_data.Numeric_Response = double(categorical(table_data.Response, ...
    {'Mostly false', 'Mostly true', 'Definitely true'}, [1,2,3]));

Quick Recommendation #

• Use categorical conversion if you need to preserve the text-numeric link for later use.
• Use logical indexing for a no-frills, fast solution with minimal code.
• Use mapping containers if you expect to expand the list of categories down the line.

内容的提问来源于stack exchange，提问作者Shosho