如何在八度中读取具有不同列数的csv文件

Question

I would like to read selected columns of string data in a csv file in which the total number of columns varies from line to line. Let's say the content of the csv file is as follows:

line 1: a1,b1,c1,d1,e1
line 2: a2,b2,c2
line 3: a3,b3,c3,d3

I would like to read the 1st column into variable a and the second column into variable b with the expectation of the following values for a and b:

a[1,1] = a1, a[2,1] = a2, a[3,1] = a3
b[1,1] = b1, b[2,1] = b2, b[3,1] = b3

I tried using textread with the format specifiers matching just the desired columns:

[a,b] = textread('SampleFile.csv','%s%s','delimiter',',')

but that generated erroneous results:

a[1,1] = a1, a[2,1] = c1, a[3,1] = e1, a[4,1] = b2, a[5,1] = b3, a[6,1] = d3
b[1,1] = b2, b[2,1] = d1, b[3,1] = a2, b[4,1] = a3, b[5,1] = c3, b[6,1] =

I also tried using textread with the format specifiers matching the maximum number of columns using dummy variables for the unwanted columns:

[a,b,dummy1,dummy2,dummy3] = textread('SampleFile.csv','%s%s%s%s%s','delimiter',',')

but that also generated erroneous results:

a[1,1] = a1, a[2,1] = a2, a[3,1] = d3
b[1,1] = b2, b[2,1] = b2, b[3,1] =

Is there a way to capture just the desired columns using either the textread command or another approach?

I am running Octave 3.8.0 on Mac OS 10.9.5.

Answer 1

Your output is expected because it's exactly mimics the example from the function documentation :

Assume a data file like:
1 a 2 b
3 c 4 d
5 e

[a, b] = textread (f, "%f %s")

returns two columns of data, one with doubles, the other a cellstr array:

a = [1; 2; 3; 4; 5]
b = {"a"; "b"; "c"; "d"; "e"}

You have to read your text file line by line to do what you want:

f = fopen ("test.txt");
c1 = {};
c2 = {};
while (l = fgetl (f)) != -1
  [a b] = strread (l, "%s %s");
  c1 = cat (1, c1, a{1});
  c2 = cat (1, c2, b{1});
endwhile
fclose (f);