Convert cell array of strings to unicode quickly
显示 更早的评论
I have an array of approximately 10M strings, and I'm interested in converting each string to its unicode values. Is there a quick, one-line way to convert the whole string array into numeric values? Ideally, I'd love a solution like this:
numeric_matrix = double(string_array);
But of course double (and unicode2native) does not support cells. So my current solution is to loop through the string array:
for ii = 1:length(string_array)
numeric_matrix(ii,:) = double(string_array{ii});
end
Unfortunately this for-loop solution is very inefficient. It can take upwards of 10 minutes for very large numbers of strings. I tried googling this but didn't see anything better. Is there a simpler, faster way to do this, ideally in one line?
采纳的回答
Walter Roberson
2016-2-2
Try
numeric_array = cellfun(@uint16, stringarray);
Try it on a smaller subset first as I do not know how the timing would compare. It should have the advantage of not needing to change the internal representation.
3 个评论
Greg
2016-2-2
Guillaume
2016-2-2
As far as I understand, matlab native encoding is not unicode but whatever is your system locale, so converting the string to double (or uint16) may not convert it to unicode unless your locale is also unicode. You would have to call native2unicode on the strings to be sure.
Most likely your cellfun is slower than a loop because you're using an anonymous function to perform your extra operation. Anonymous function calls have a significant overhead in matlab.
Greg
2016-2-2
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Data Type Conversion 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
