package base4;
import tool.*;
public class pro {
int s,datanumber;
String[] data=new String[5000];
int[] a=new int[79];
int[] b=new int[79];
String[] k=new String[5000];
public static void main(String[] args) {
pro t=new pro();
}
pro(){
int tr;
for(tr=1990;tr<2016;tr++){
String urlx="https://ja.wikipedia.org/wiki/"+tr+"%E5%B9%B4%E5%BA%A6%E6%96%B0%E4%BA%BA%E9%81%B8%E6%89%8B%E9%81%B8%E6%8A%9E%E4%BC%9A%E8%AD%B0_(%E6%97%A5%E6%9C%AC%E3%83%97%E3%83%AD%E9%87%8E%E7%90%83)";
getdata sub=new getdata();
sub.makedata(urlx,"UTF-8");
data=sub.data;
datanumber=sub.datanumber;
int s1=0;
for(s=1;s<datanumber+1;s++){
int h=0;
if(data[s].indexOf("指名")>-1)h=h+1;
if(data[s].indexOf("<h2>")>-1)h=h+1;
if(data[s].indexOf("重複")>-1)h=0;
if(h==2)s1=s;
}
int s2=0;
for(s=s1+1;s<datanumber+1;s++){
int h=0;
if(data[s].indexOf("<h2>")>-1)h=h+1;
if(s2==0)h=h+1;
if(h==2)s2=s;
}
int sx=0;
for(s=s1;s<s2;s++){
if(data[s].indexOf("<table")>-1)sx=sx+1;
if(data[s].indexOf("<table")>-1)a[sx]=s;
}
int number=sx;
sx=0;
for(s=s1;s<s2;s++){
if(data[s].indexOf("</table")>-1)sx=sx+1;
if(data[s].indexOf("</table")>-1)b[sx]=s;
}
for(s=1;s<number+1;s++){
String str="";
for(sx=a[s];sx<b[s]+1;sx++){
str=str+data[sx];
}
k[s]=str;
}
int n;
String[] z=new String[10];
for(s=1;s<number+1;s++){
String[] x=k[s].split("</tr>");
for(sx=1;sx<x.length;sx++){
x[sx]=x[sx].replace("th","td");
String[] y=x[sx].split("</td>");
z[0]="";
z[1]="";
z[2]="";
z[3]="";
z[4]="";
for(n=0;n<y.length;n++){
y[n]=y[n].replace("<td>","");
y[n]=y[n].replace("<tr>","");
y[n]=y[n].replace("</table>","");
z[n]=y[n];
}
System.out.println(tr+","+s+","+z[0]+","+z[1]+","+z[2]+","+z[3]+","+z[4]);
}
}
}
}
}
最終更新:2016年04月10日 10:12