How to rename identical variables under one common name?

I've read in a Excel file through MATLAB that is a 1191x12 Table.

The goals is to rename identical variables to one common variable name, Example below:

For any type that start with ABCD, Id like to rename to only ABCD removing the following characters/numbers.

For any type that starts with BACA, Id like to rename to only BACA removing the following characters/numbers.

For any type that starts with CABD, Id like to rename to only CABD removing the following characters/numbers.

For any type that starts with DABC, Id like to rename to only CABD removing the following characters/numbers.

Current Table Example below:

Expected Table Below:

[Stephen23]

Not enough information: do you want to split/delete those characters based on their content/character type (e.g. digits) or based on their location in the text (e.g. indices >4) or based on some other condition (e.g. what?)...?

Do you know in advance the "common names", or do you want to automatically detect the "common names" ?

Please upload your data file by clicking the paperclip button.

[Dyuman Joshi]

If all the data is 4 letters followed by 3 digits, simply remove the digits by indexing or store the 1st 4 values only.

[Matt J]

It is better to attach your examples as .mat files with actual data variables in them.

[Rookie Programmer]

Thanks, the code below worked for this issue.

numrows = height(Table)

NewType = table

NewType.Types = cell(numrows,1)

for IdxR = 1:numrows

if contains(table.type{IdxR}, 'ABCD')

NewType.Types{IdxR} = 'ABCD'

elseif contains(table.type{IdxR}, 'BACA')

NewType.Types{IdxR} = 'BACA'

elseif contains(table.type{IdxR}, 'CABD')

NewType.Types{IdxR} = 'CABD'

elseif contains(table.type{IdxR}, 'DABC')

NewType.Types{IdxR} = 'DABC'

end

end

Table.Type = NewType.Types;

Answer 1

[Stephen23]

T = readtable('myData.xlsx')

T = 4×3 table
    Place     Day       Type    
    ______    ___    ___________

    {'NC'}     1     {'ABCD123'}
    {'NY'}     2     {'BACA654'}
    {'TX'}     3     {'CABD154'}
    {'WV'}     4     {'DABC987'}

T.Type = regexp(T.Type,'^[A-Z]{4}','match','once')

T = 4×3 table
    Place     Day      Type  
    ______    ___    ________

    {'NC'}     1     {'ABCD'}
    {'NY'}     2     {'BACA'}
    {'TX'}     3     {'CABD'}
    {'WV'}     4     {'DABC'}

[Jon]

Great that regexp operates on whole cell array so no need to use cellfun.

On the other hand you seem to only look at the first 4 elements of the string. Your code won't strip of the numerical part of the strings if there are more than 4 leading characters as shown below for the 3rd row.

T = readtable('myData2.xlsx')

T = 4×3 table
    Place     Day        Type    
    ______    ___    ____________

    {'NC'}     1     {'ABCD123' }
    {'NY'}     2     {'BACA654' }
    {'TX'}     3     {'CABDE154'}
    {'WV'}     4     {'DABC987' }

T.Type = regexp(T.Type,'^[A-Z]{4}','match','once')

T = 4×3 table
    Place     Day      Type  
    ______    ___    ________

    {'NC'}     1     {'ABCD'}
    {'NY'}     2     {'BACA'}
    {'TX'}     3     {'CABD'}
    {'WV'}     4     {'DABC'}

If you already know that you just want the first 4 characters, no need to use regexp, or cell fun, just use extractBetween

T = readtable('myData.xlsx')

T = 4×3 table
    Place     Day       Type    
    ______    ___    ___________

    {'NC'}     1     {'ABCD123'}
    {'NY'}     2     {'BACA654'}
    {'TX'}     3     {'CABD154'}
    {'WV'}     4     {'DABC987'}

T.Type = extractBetween(T.Type,1,4)

T = 4×3 table
    Place     Day      Type  
    ______    ___    ________

    {'NC'}     1     {'ABCD'}
    {'NY'}     2     {'BACA'}
    {'TX'}     3     {'CABD'}
    {'WV'}     4     {'DABC'}

[Stephen23]

"On the other hand you seem to only look at the first 4 elements of the string."

That is exactly why I asked the OP for clarification, what their specific requirements are:

https://www.mathworks.com/matlabcentral/answers/1992643-how-to-rename-identical-variables-under-one-common-name#comment_2807218

So far the OP has not stated how many leading characters they want to retain, we only have their examples to go by: are they complete and representative? Are they always uppercase? I doubt it... but no one here knows.

"Your code won't strip of the numerical part of the strings if there are more than 4 leading characters as shown below for the 3rd row."

The OP states that they wish to "removing the following characters/numbers", so your assumption that the characters to remove are "numerical parts" seems to be inconsistent with what the OP states (if they only mean "numerical parts" as you wrote what do they mean by "characters"?). But again, this is why I asked for clarification. I see little point in developing and testing regular expressions until I have a reasonably clear specification.

"If you already know that you just want the first 4 characters, no need to use regexp, or cell fun, just use extractBetween"

Use https://www.mathworks.com/help/matlab/ref/extractbefore.html

Answer 2

[sushma swaraj]

Hi, You can use the 'startsWith' function instead of 'contains' to get the appropriate result.

https://www.mathworks.com/help/matlab/ref/startswith.html?s_tid=doc_ta

https://www.mathworks.com/help/matlab/ref/string.contains.html?searchHighlight=contains&s_tid=srchtitle_contains_1

Hope it helps you in modifying your code!

Answer 3

[Jon]

Here's a general way to handle this. Note no need for all those if statements

% Read in the data
T = readtable('myData.xlsx'); % put the name of your data file here

% Strip off the numeric part of the data in the 'Type' column
T.Type = cellfun(@(x) x(~isstrprop(x,'digit')),T.Type,'UniformOutput',false)


% Save the data back into a new Excel file (not sure if you need to do
% this)
writetable(T,'myData_stripped.xlsx')

[Jon]

Explaining a little further

The function isstrprop(x,'digit') returns a logical vector of 1's and zeros with a 1 (true) wherever x contains a numeric digit

isstrprop('ABCD123','digit')

ans = 1×7 logical array
   0   0   0   0   1   1   1

So notting this, e.g. ~isstrprop('ABCD123','digit') will give you a true wherever the name has a nonnumeric letter

~isstrprop('ABCD123','digit')

ans = 1×7 logical array
   1   1   1   1   0   0   0

If we just want to keep the non-numeric elements in the string we can use logical indexing to select them

name = 'ABCD123'

name = 'ABCD123'

newName = name(~isstrprop(name,'digit'))

newName = 'ABCD'

In your case you want to do this for a whole column in a table. The column in the table is actually a m by 1 cell array of values, that would look something like this

names = {'ABCD123';'BACCA654';'CABD154'}

names = 3×1 cell array
    {'ABCD123' }
    {'BACCA654'}
    {'CABD154' }

To do something to every element in a cell array you use cellfun, so putting it all together using the anonymous function @(x) x(~isstrprop(x,'digit')) which says what to do with each element x

newNames = cellfun(@(x) x(~isstrprop(x,'digit')),names,'UniformOutput',false)

newNames = 3×1 cell array
    {'ABCD' }
    {'BACCA'}
    {'CABD' }

You can read more about anonymous functions here https://www.mathworks.com/help/matlab/matlab_prog/anonymous-functions.html

and about cellfun here https://www.mathworks.com/help/matlab/ref/cellfun.html

[Jon]

As noted in my comment to @Stephen23

If you already know that you just want the first 4 characters, no need to use regexp, or cellfun, just use extractBetween

T = readtable('myData.xlsx')

T = 4×3 table
    Place     Day       Type    
    ______    ___    ___________

    {'NC'}     1     {'ABCD123'}
    {'NY'}     2     {'BACA654'}
    {'TX'}     3     {'CABD154'}
    {'WV'}     4     {'DABC987'}

T.Type = extractBetween(T.Type,1,4)

T = 4×3 table
    Place     Day      Type  
    ______    ___    ________

    {'NC'}     1     {'ABCD'}
    {'NY'}     2     {'BACA'}
    {'TX'}     3     {'CABD'}
    {'WV'}     4     {'DABC'}

[Jon]

Did any of our responses answer your question? If so please accept an answer so that others will know an answer is available. If not please let us know what aspect of your problem we are missing.

Answer 4

[Peter Perkins]

This just SCREAMS for using categorical. Screams. This might seem like more work, but untimately you will be happier.

TextData = ["aa";"bb";"cc";"aa";"bb";"cc";"aa";"bb";"cc";"aa"] + randi([100 200],10,1);
t = table(rand(10,1),TextData)

t = 10×2 table
      Var1      TextData
    ________    ________

     0.22574    "aa165" 
     0.32424    "bb170" 
    0.064198    "cc157" 
     0.39656    "aa124" 
     0.49629    "bb123" 
     0.55708    "cc191" 
     0.16259    "aa152" 
     0.36202    "bb116" 
      0.7459    "cc108" 
     0.68575    "aa125" 

t.CatData = categorical(t.TextData)

t = 10×3 table
      Var1      TextData    CatData
    ________    ________    _______

     0.22574    "aa165"      aa165 
     0.32424    "bb170"      bb170 
    0.064198    "cc157"      cc157 
     0.39656    "aa124"      aa124 
     0.49629    "bb123"      bb123 
     0.55708    "cc191"      cc191 
     0.16259    "aa152"      aa152 
     0.36202    "bb116"      bb116 
      0.7459    "cc108"      cc108 
     0.68575    "aa125"      aa125 

oldCats = string(categories(t.CatData));
newCats = unique(extractBetween(oldCats,1,2))

newCats = 3×1 string array
    "aa"
    "bb"
    "cc"

for i = 1:length(newCats)
    toBeMerged = startsWith(oldCats,newCats(i));
    t.CatData = mergecats(t.CatData,oldCats(toBeMerged),newCats(i));
end
t.TextData = []

t = 10×2 table
      Var1      CatData
    ________    _______

     0.22574      aa   
     0.32424      bb   
    0.064198      cc   
     0.39656      aa   
     0.49629      bb   
     0.55708      cc   
     0.16259      aa   
     0.36202      bb   
      0.7459      cc   
     0.68575      aa