expecting EOF, found 'columns'..cant figure it out

Programming Questions

sha89red

expecting EOF, found 'columns'..cant figure it out

in Programming Questions • 11 months ago

Could somebody help me solving the error??
I couldn't figure it out..
The 1st code will run without any problem..

String rows[] = loadStrings("Head.txt"); //read txt files
String[] columns;
columns = splitTokens(rows[0]); //how many columns in a line in txt file
double[][] matrixA = new double[rows.length][columns.length];
double[][] transposed_matrixA = new double[columns.length][rows.length];
double[][] covariance_matrixA = new double[columns.length][columns.length];
double[] Mean = new double[columns.length];
double[][] X = new double[columns.length][rows.length];
double[][] transposed_X = new double[rows.length][columns.length];
//double[][] V = new double[columns.length][columns.length];
//double[][] D = new double[columns.length][columns.length];
//check the size of input matrix from the .txt file
println("there are " + rows.length + " rows");
println("there are " + columns.length + " columns");
//insert the numbers from txt file into matrixA
for (int i =0 ; i<rows.length; i++){
columns = splitTokens(rows[i]);
for(int j=0; j<columns.length; j++){
matrixA[i][j] = Double.parseDouble(columns[j]);
transposed_matrixA[j][i] = matrixA[i][j];
}
}
//calculate means for transposed_matrixA.So we have to change rows to columns and vice versa
for(int i=0; i<columns.length; i++){
double mean = 0.0;
for(int j=0; j<rows.length; j++){
mean += transposed_matrixA[i][j];
}
Mean[i] = mean/rows.length;
}
for(int i=0; i<columns.length; i++){
for(int j=0; j<rows.length; j++){
X[i][j] = transposed_matrixA[i][j] - Mean[i];
transposed_X[j][i] = X[i][j];
}
}
for(int i=0; i<columns.length; i++){
for(int j=0; j<columns.length; j++){
for(int k=0; k<rows.length; k++){
covariance_matrixA[i][j] += X[i][k]*transposed_X[k][j];
}
covariance_matrixA[i][j] /= rows.length;
}
}

But when I add a code as below whether in a new tab or below line 50 of above code, I got an error stating " expecting EOF, found 'columns' " that is highlighted at the 3rd line of above code..The weird part is,even i commented out below's code, the error still remains..only when I erase the code that the error will gone..
Help please

class jacobi{
static final double EPS = pow(10,-6);
static final double TINY = pow(10,-20);
static final int MAX=ITER = 100;
static void kaiten(double[][] a, int i, int j, int k, int l, double c, double s){
double x = a[i][j], y = a[k][l];
a[i][j] = x*c - y*s;
a[k][l] = x*s + y*c;
}
public static boolean solve(double[][] a, double[][] w){
int n = a.length;
double s = 0, offdiag = 0;
for(int j=0; j<n; j++){
for(int k=0; k<n; k++){
w[j][k] = 0;
}
w[j][j] = 1; s += a[j][j] * a[j][j];
for(int k = j+1; k<n; k++){
offdiag += a[j][k];
}
}
double tolerance = EPS * EPS * (s/2 + offdiag);
int iter;
for(iter=1; iter<=MAX_ITER; iter++){
offdiag = 0;
for(int j=0; j<n-1; j++){
for(int k=j+1; k<n; k++){
offdiag += a[j][k] * a[j][k];
}
println(iter + " : " + sqrt(2*offdiag / (n*(n-1))));
}
if(offdiag<tolerance){
break;
}
for(int j=0; j<n-1; j++){
for(int k=j+1; k<n; k++){
if(abs(a[j][k]) < TINY){
continue;
}
double t = (a[k][k] - a[j][j]) / (2 * a[j][k]);
if(t >= 0){
t = 1 /(t + sqrt(t*t + 1));
}
else{
t = 1/(t - sqrt(t*t + 1));
}
double c = 1/sqrt(t*t + 1), u = t*c;
t *= a[j][k];
a[j][j] -= t; a[k][k] += t; a[j][k] = 0;
int i;
for(i=0; i<j; i++) {
kaiten(a,i,j,i,k,c,u);
}
for(i=j+1; i<k; i++) {
kaiten(a,j,i,i,k,c,u);
}
for(i=k+1; i<n; i++) {
kaiten(a,j,i,k,i,c,u);
}
for(i=0; i<n; i++) {
kaiten(w,j,i,k,i,c,u);
}
}
}
}
if(iter > MAX_ITER) return false;
for(int i=0; i<n-1; i++){
int k=i; double t=a[k][k];
for(int j=i+1; j<n; j++){
if(a[j][j] > t){
k=j; t=a[k][k];
}
}
a[k][k] = a[i][i]; a[i][i] = t;
double[] v = w[k]; w[k] = w[i]; w[i] = v;
}
return true;
}
}

Replies(6)

v.k.

Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

I think that is because when you do not have a class, Processing will wrap all your code in a setup() method, but with a class mixed in it get lost about that, and the split needs to be done in setup or draw. If you add a setup() it will work, or at least complain of other things... check it out:

String rows[] = loadStrings("Head.txt"); //read txt files
String[] columns;
//how many columns in a line in txt file
double[][] matrixA = new double[rows.length][columns.length];
double[][] transposed_matrixA = new double[columns.length][rows.length];
double[][] covariance_matrixA = new double[columns.length][columns.length];
double[] Mean = new double[columns.length];
double[][] X = new double[columns.length][rows.length];
double[][] transposed_X = new double[rows.length][columns.length];
//double[][] V = new double[columns.length][columns.length];
//double[][] D = new double[columns.length][columns.length];
void setup()
{
columns = splitTokens(rows[0]);
//check the size of input matrix from the .txt file
println("there are " + rows.length + " rows");
println("there are " + columns.length + " columns");
//insert the numbers from txt file into matrixA
for (int i =0 ; i<rows.length; i++){
columns = splitTokens(rows[i]);
for(int j=0; j<columns.length; j++){
matrixA[i][j] = Double.parseDouble(columns[j]);
transposed_matrixA[j][i] = matrixA[i][j];
}
}
//calculate means for transposed_matrixA.So we have to change rows to columns and vice versa
for(int i=0; i<columns.length; i++){
double mean = 0.0;
for(int j=0; j<rows.length; j++){
mean += transposed_matrixA[i][j];
}
Mean[i] = mean/rows.length;
}
for(int i=0; i<columns.length; i++){
for(int j=0; j<rows.length; j++){
X[i][j] = transposed_matrixA[i][j] - Mean[i];
transposed_X[j][i] = X[i][j];
}
}
for(int i=0; i<columns.length; i++){
for(int j=0; j<columns.length; j++){
for(int k=0; k<rows.length; k++){
covariance_matrixA[i][j] += X[i][k]*transposed_X[k][j];
}
covariance_matrixA[i][j] /= rows.length;
}
}
}
class jacobi{
static final double EPS = pow(10,-6);
static final double TINY = pow(10,-20);
static final int MAX=ITER = 100;
static void kaiten(double[][] a, int i, int j, int k, int l, double c, double s){
double x = a[i][j], y = a[k][l];
a[i][j] = x*c - y*s;
a[k][l] = x*s + y*c;
}
public static boolean solve(double[][] a, double[][] w){
int n = a.length;
double s = 0, offdiag = 0;
for(int j=0; j<n; j++){
for(int k=0; k<n; k++){
w[j][k] = 0;
}
w[j][j] = 1; s += a[j][j] * a[j][j];
for(int k = j+1; k<n; k++){
offdiag += a[j][k];
}
}
double tolerance = EPS * EPS * (s/2 + offdiag);
int iter;
for(iter=1; iter<=MAX_ITER; iter++){
offdiag = 0;
for(int j=0; j<n-1; j++){
for(int k=j+1; k<n; k++){
offdiag += a[j][k] * a[j][k];
}
println(iter + " : " + sqrt(2*offdiag / (n*(n-1))));
}
if(offdiag<tolerance){
break;
}
for(int j=0; j<n-1; j++){
for(int k=j+1; k<n; k++){
if(abs(a[j][k]) < TINY){
continue;
}
double t = (a[k][k] - a[j][j]) / (2 * a[j][k]);
if(t >= 0){
t = 1 /(t + sqrt(t*t + 1));
}
else{
t = 1/(t - sqrt(t*t + 1));
}
double c = 1/sqrt(t*t + 1), u = t*c;
t *= a[j][k];
a[j][j] -= t; a[k][k] += t; a[j][k] = 0;
int i;
for(i=0; i<j; i++) {
kaiten(a,i,j,i,k,c,u);
}
for(i=j+1; i<k; i++) {
kaiten(a,j,i,i,k,c,u);
}
for(i=k+1; i<n; i++) {
kaiten(a,j,i,k,i,c,u);
}
for(i=0; i<n; i++) {
kaiten(w,j,i,k,i,c,u);
}
}
}
}
if(iter > MAX_ITER) return false;
for(int i=0; i<n-1; i++){
int k=i; double t=a[k][k];
for(int j=i+1; j<n; j++){
if(a[j][j] > t){
k=j; t=a[k][k];
}
}
a[k][k] = a[i][i]; a[i][i] = t;
double[] v = w[k]; w[k] = w[i]; w[i] = v;
}
return true;
}
}

sha89red

Re: Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

got it..thanks v.k.
I think its just like you said..other than class i used other functions like void and boolean in it..
looks like if I create such function or class..the void setup() or maybe draw() becomes necessary..

in line 5~13 we declared array of doubles using columns.length which wasn't defined before
and we only defined the columns in line 18 in setup() after..

it leads to an "java.lang.NullPointer.Exception" error..

any ideas to declare the variables at global scope??

GoToLoop

Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

About the other things, I've got some to say:

Processing is not so fond of the data-type double.
If you don't need such high precision, I'd advise you to use float instead.
If you need double, you have to add Math. in front of most calculation functions like:
sqrt(); -> Math.sqrt();

sha89red

Re: Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

many thanks for the advice..
I need high precision as im doing recognition system..and its just as you said,using sqrt() or abs() needs to be in float type so when using double, have to use Math.sqrt() Math.abs()..
I just cast it with (float) but there would be no meaning to use double then..So im thinking of changing it to Math. then..

I have questions..
1.would using double data-type effecting processing in some way or its fine using it without any problem??
2.Do i need to import a library in order to use Math. ??

Sorry for noob questions too..

Again thank you..

GoToLoop

Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

Raw Java is a little intimidating for casual programmers like us.
The beauty of Processing is that it tries to make Java more palatable.

One of the strategies it uses is to wrap up the original Java methods w/ something more clean & easy.
Like, instead of Math.sqrt(), we just need to type sqrt() w/o the Math.!

Processing creators rationalized that most people who would program in it would rarely have the need for double precision calculations!
So most of those clean & neat wrap-ups from the original Java methods would only use either int or float!
And pretty much that's the standard for Processing!

But since we still have access of the original Java methods, we can use them when the need arises, like your case!

So there's not much problem, unless you wanna also convert your code into processing.js, which is Processing under JavaScript!

And, about importing any Java libraries, most of them are automatically imported by the Processing's preprocessor anyways!

sha89red

Re: Re: expecting EOF, found 'columns'..cant figure it out

11 months ago

Thanks for the explanation..I appreciate it..

Top Reply