How can I please reduce the code running time?

3 vues (au cours des 30 derniers jours)
MAHMOUD ALZIOUD
MAHMOUD ALZIOUD le 9 Nov 2017
Réponse apportée : Jan le 10 Nov 2017
Dear All, I wrote this code to analyze my text data, each time I run this code it takes exactly 22 minutes to finish ! which is very long time, since I am new on MATLAB I still don't know the tricks to minimize running time, will any body please give me some ideas regarding this code? Thank you MATLAB Community
clear;
clc;
directory=dir('*.Y07');
count=0;
for K = 1 : length(directory)
filename = directory(K).name;
fileID = fopen(filename,'r');
formatSpec = '%s';
A_cell = textscan(fileID,formatSpec);
A=char(A_cell{1,1}{:,:});
A(find(isnan(A)))=0;
[rows,columns]=size(A);
if columns~=105
ArrayTemp=zeros(rows,105);
ArrayTemp(1:rows,1:columns)=A;
A=ArrayTemp;
A=char(A);
A(isspace(A)) = '0';
end
x1=filename;
xtr=strcat('C:\Users\maa285\Desktop\New folder (2)\',x1);
fid = fopen( xtr, 'wt' );
Record_Type = A(:,1:1);
FibsCode = A(:,2:3);
StationID = A(:,4:9);
Direction_Of_Travel = A(:,10:10);
Lane_Of_Travel = A(:,11:11);
Year_of_Data = A(:,12:13);
Month_of_Data = A(:,14:15);
Day_of_Data = A(:,16:17);
Hour_of_Data = A(:,18:19);
Vehicle_Class = A(:,20:21);
Open = A(:,22:24);
Total_Weight_of_vehicle = A(:,25:28);
Number_of_axles = A(:,29:30);
A_Axle_Weight = A(:,31:33);
A_B_Axle_spacing = A(:,34:36);
B_Axle_Weight = A(:,37:39);
B_C_Axle_spacing = A(:,40:42);
C_Axle_Weight = A(:,43:45);
C_D_Axle_spacing = A(:,46:48);
D_Axle_Weight = A(:,49:51);
D_E_Axle_spacing = A(:,52:54);
E_Axle_Weight = A(:,55:57);
E_F_Axle_spacing = A(:,58:60);
F_Axle_Weight = A(:,61:63);
F_G_Axle_spacing= A(:,64:66);
G_Axle_Weight = A(:,67:69);
G_H_Axle_spacing = A(:,70:72);
H_Axle_Weight = A(:,73:75);
H_I_Axle_spacing = A(:,76:78);
I_Axle_Weight= A(:,79:81);
I_J_Axle_spacing = A(:,82:84);
J_Axle_Weight = A(:,85:87);
J_K_Axle_spacing= A(:,88:90);
K_Axle_Weight = A(:,91:93);
K_L_Axle_spacing = A(:,94:96);
L_Axle_Weight = A(:,97:99);
L_M_Axle_spacing = A(:,100:102);
M_Axle_Weight = A(:,103:105);
%This is to convert the string to numbers so it can show the output:
Ans_1=str2num(Record_Type);
fprintf(fid,Record_Type);
Ans_2=str2num(FibsCode);
fprintf(fid,FibsCode);
Ans_3=str2num(StationID);
fprintf(fid,StationID);
Ans_4=str2num(Direction_Of_Travel);
fprintf(fid,Direction_Of_Travel);
Ans_5=str2num(Lane_Of_Travel);
fprintf(fid,Lane_Of_Travel);
Ans_6=str2num(Year_of_Data);
fprintf(fid,Year_of_Data);
Ans_7=str2num(Month_of_Data);
fprintf(fid,Month_of_Data);
Ans_8=str2num(Day_of_Data);
fprintf(fid,Day_of_Data);
Ans_9=str2num(Hour_of_Data);
fprintf(fid,Hour_of_Data);
Ans_10=str2num(Vehicle_Class);
fprintf(fid,Vehicle_Class);
Ans_11=str2num(Open);
fprintf(fid,Open);
Ans_12=str2num(Total_Weight_of_vehicle);
fprintf(fid,Total_Weight_of_vehicle);
Ans_13=str2num(Number_of_axles);
fprintf(fid,Number_of_axles);
Ans_14=str2num(A_Axle_Weight);
fprintf(fid,A_Axle_Weight);
Ans_15=str2num(A_B_Axle_spacing);
fprintf(fid,A_B_Axle_spacing);
Ans_16=str2num(B_Axle_Weight);
fprintf(fid,B_Axle_Weight);
Ans_17=str2num(B_C_Axle_spacing);
fprintf(fid,B_C_Axle_spacing);
Ans_18=str2num(C_Axle_Weight);
fprintf(fid,C_Axle_Weight);
Ans_19=str2num(C_D_Axle_spacing);
fprintf(fid,C_D_Axle_spacing);
Ans_20=str2num(D_Axle_Weight);
fprintf(fid,D_Axle_Weight);
Ans_21=str2num(D_E_Axle_spacing);
fprintf(fid,D_E_Axle_spacing);
Ans_22=str2num(E_Axle_Weight);
fprintf(fid,E_Axle_Weight);
Ans_23=str2num(E_F_Axle_spacing);
fprintf(fid,E_F_Axle_spacing);
Ans_24=str2num(F_Axle_Weight);
fprintf(fid,F_Axle_Weight);
Ans_25=str2num(F_G_Axle_spacing);
fprintf(fid,F_G_Axle_spacing);
Ans_26=str2num(G_Axle_Weight);
fprintf(fid,G_Axle_Weight);
Ans_27=str2num(G_H_Axle_spacing);
fprintf(fid,G_H_Axle_spacing);
Ans_28=str2num(H_Axle_Weight);
fprintf(fid,H_Axle_Weight);
Ans_29=str2num(H_I_Axle_spacing);
fprintf(fid,H_I_Axle_spacing);
Ans_30=str2num(I_Axle_Weight);
fprintf(fid,I_Axle_Weight);
Ans_31=str2num(I_J_Axle_spacing);
fprintf(fid,I_J_Axle_spacing);
Ans_32=str2num(J_Axle_Weight);
fprintf(fid,J_Axle_Weight);
Ans_33=str2num(J_K_Axle_spacing);
fprintf(fid,J_K_Axle_spacing);
Ans_34=str2num(K_Axle_Weight);
fprintf(fid,K_Axle_Weight);
Ans_35=str2num(K_L_Axle_spacing);
fprintf(fid,K_L_Axle_spacing);
Ans_36=str2num(L_Axle_Weight);
fprintf(fid,L_Axle_Weight);
Ans_37=str2num(L_M_Axle_spacing);
fprintf(fid,L_M_Axle_spacing);
Ans_38=str2num(M_Axle_Weight);
fprintf(fid,M_Axle_Weight);
%to establish a new arrays for the date:
monthvar = Month_of_Data;
dayvar = Day_of_Data;
yearvar = Year_of_Data;
datevar = strcat(num2str(monthvar),'/',num2str(dayvar),'/',num2str(yearvar));
% DayNumber = weekday(datevar);
[DayNumber,DayName] = weekday(datevar);
%%to return the separated vectors into one new matrix with usable data:
all=[Ans_2,Ans_2,Ans_3,Ans_4,Ans_5,Ans_6,Ans_7,Ans_8,Ans_9,Ans_10,DayNumber,Ans_12,Ans_13,Ans_14,Ans_15,Ans_16,Ans_17,Ans_18,Ans_19,Ans_20,Ans_21,Ans_22,Ans_23,Ans_24,Ans_25,Ans_26,Ans_27,Ans_28,Ans_29,Ans_30,Ans_31,Ans_32,Ans_33,Ans_34,Ans_35,Ans_36,Ans_37,Ans_38];
% all(isspace(all)) = '0';
%now select each class data out of the original data A.
VehClass = Ans_10;
Class9_data = all(VehClass == 9, :);
count=count+1;
end
  1 commentaire
MAHMOUD ALZIOUD
MAHMOUD ALZIOUD le 9 Nov 2017
Modifié(e) : MAHMOUD ALZIOUD le 9 Nov 2017
Note, this code is used to read multiple files, that is why I use the fprintf command, I have K numbers of files.

Connectez-vous pour commenter.

Réponse acceptée

Walter Roberson
Walter Roberson le 9 Nov 2017
You are using fprintf() with no format specifier, passing in what looks to be character strings. The character strings are going to be interpreted as being format strings. That will mostly result in the character strings being output. However, they will be output with no spaces between them and no delimiter.
You can improve performance by calling fprintf() fewer times, such as
fprintf(fid, '%s %s %s %s %s', H_Axle_Weight, H_I_Axle_Weight, I_Axle_Weight, I_J_Axle_Weight)
  1 commentaire
MAHMOUD ALZIOUD
MAHMOUD ALZIOUD le 9 Nov 2017
Thank you very much Mr Walter, I will modify my code and run it now.

Connectez-vous pour commenter.

Plus de réponses (2)

Gregory McFadden
Gregory McFadden le 9 Nov 2017
try running your code after issuing
profile on
and then when it is done use
profile viewer
that will tell you exactly what line(s) are consuming all the time in the code, then we can focus on those specific long compute time issues
  5 commentaires
MAHMOUD ALZIOUD
MAHMOUD ALZIOUD le 9 Nov 2017
Thank you very much for your advice, I will try it now
MAHMOUD ALZIOUD
MAHMOUD ALZIOUD le 9 Nov 2017
this saved me 5 minutes, instead of 22 minutes per run, now it is 17. so it worked. thank you for this

Connectez-vous pour commenter.


Jan
Jan le 10 Nov 2017
After
A=char(A_cell{1,1}{:,:});
A is a CHAR. Then it cannot contain NaNs, because this can happen for DOUBLE and SINGLE only. The line:
A(find(isnan(A)))=0;
is useless in consequence.
The numbered variables "Ans_1" look cruel. It is a horror to debug this.

Catégories

En savoir plus sur Matrix Indexing dans Help Center et File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by