Read csv data directly from url in Matlab

Question

I am having some problems reading data from a database in Matlab. The url link I use to download the data gives a semicolon delimited text file and I need Matlab to identify this data and arrange it accordingly in a struct format (since the data has different classes) for example. I have already used urlread and could download the data succesfully, the only problem is I am getting all the data as a character string inside one cell and I need to get this data as a table and well organized.

Basically I would like to know if it is possible to load data from a url into Matlab the same way the read.csv function in R does, where you just put the url where the file name should go and define how the data is delimited and viola, you get your data.frame with all your data perfectly organized as it should.

I suppose there are ways to interpret the character string after using urlread and convert it somehow into an organized struct variable but there has to be a way to read it directly from the url as R does.

Answer 1

Here is a piece of code that will read the csv data from the web ( urlread ), use textscan to scan and format the data into cells (strings and scalars allowed), then convert the cells into a structure with cell2struct . The structure created keeps the textscan formatting.

Note that you have to define the textscan format and the cell2struct input to suit your data.

block = urlread('http://hci.stanford.edu/jheer/workshop/data/florida2000/Florida2000.csv');
C = textscan(block,'%s%s%f%s%f','HeaderLines',1,'EndOfLine','\n');
S = cell2struct(C,{'county','technology','columns','category','ballots'},2)

Here is the Florida 2000 Presidential Election Results ( .csv , 938 data points)

county,technology,columns,category,ballots
Alachua,Optical,1,under,217
Alachua,Optical,1,over,105
Alachua,Optical,1,Bush,34124
Alachua,Optical,1,Gore,47365
Alachua,Optical,1,Browne,658
Alachua,Optical,1,Nader,3226
Alachua,Optical,1,Harris,6
...

that will produce

S = 

    county: {938x1 cell}    %string
technology: {938x1 cell}    %string
   columns: [938x1 double]  %double
  category: {938x1 cell}    %string
   ballots: [938x1 double]  %double

---

Edit

For double quoted text, you can use %q instead of %s in calling textscan ( FormatSpec options ) just like that

C = textscan(fileID,'%q%f');

Answer 2

Look into a function called dlmread . This will allow you to put in a string of data, tell it what the delimiter is, and it should pump out what you need.

dlmread

results = dlmread('http://someurl.com/somefile.txt',';')