MATLAB Answers

Problem with using readtable:

1 view (last 30 days)
Here is the code:
opts = detectImportOptions("C:\Users\Onat\Desktop\392\vocab.txt");
opts.VariableTypes=["string", "double"];
opts.LineEnding = ["\n"];
vocab = readtable('C:\Users\Onat\Desktop\392\vocab.txt',opts);
I'am working on an NLP application in which I need the vocabulary and frequency of those words in vocabulary. Naturally, the corpus contains tokens such as single apostrophe. It seems that this is a major problem for MATLAB since it detects it as a special char. Notice that in the output given below, after apostrophe the frequencies are seen as comments to MATLAB. Can anyone help with this issue?
vocab.txt is as follows:
..... (i.e. this is not the beginning)
and 699333
in 603607
" 538122
to 504540
a 476836
was 304423
...... (i.e continues)
the output is as follows:
...... (i.e. this is not the beginning)
"and" 6.9933e+05
"in" 6.0361e+05
" 538122↵to 504540↵a 476836↵was 304423↵The 246510↵- 229901↵is 225721↵for 198733↵)

Accepted Answer

Aghamarsh Varanasi
Aghamarsh Varanasi on 21 Apr 2021
Hi,
The quotes in a text file are used to import strings that contain any charater (even with delimiters). That is why you are seeing this behavior. You can specify the 'Format' of the file you are importing by using the 'Format' Name-Value pair input to the function readtable.
Example:
vocab = readtable('data.txt', 'Format' ,'%s %f');
This would import the data as expected.
For more info on how strings are imported from text documents, refer to this Community post.

More Answers (1)

Cris LaPierre
Cris LaPierre on 21 Apr 2021
I would take a different approach. I would use readlines and string manipulation to create the table.
str = readlines("vocab.txt")
str = 6×1 string array
"and 699333" "in 603607" "" 538122" "to 504540" "a 476836" "was 304423"
T = array2table(split(str),'VariableNames',["Word","Freq"]);
T.Freq = str2double(T.Freq)
T = 6×2 table
Word Freq _____ __________ "and" 6.9933e+05 "in" 6.0361e+05 """ 5.3812e+05 "to" 5.0454e+05 "a" 4.7684e+05 "was" 3.0442e+05

Products

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!