象这种用加号定义一个String变量是否效率很低，我的意思是这里会不会生成多个中间String变量

相应的"+"用 append方法替代！

楼上的说的有道理，因为，你在用”+“操作时，因为string类型是不可变的，所以，你没”+“一次，在物理上都是重新开辟内存的，所以效率自然第，而stringbuffer则不是，他是可变字符串，你可以用append进行追加。他始终用一块内存空间。

我见到的书都是这样说的String 的+ 操作效率较低，如果可能应该用StringBuffer
但可能需要考虑代码的可读性
如果是不变的，可以考虑放到资源文件里

谢谢大家参与，下午我写了个程序验证，发现大家都错了。
原先我是想定义一个sql文，为了好看，就写成分行的形式，找不到换行符，后来听人说定义时加号就是换行符，也就是说定义String时与非定义的语句不同。看看下面的例子：
import java.util.Date;class  TTT
{
public static void main(String[] args)
{
Date before = new Date();
long a,b;
a = before.getTime();
StringBuffer strBuff = new StringBuffer(1280000); for (int i=0; i<1000; i++ )
{
  final String SQL =
        "SELECT \n"+
        "TRIM(A.FML_CUST_NM) FML_CUST_NM,\n"+
        "TRIM(A.CAI_BANK_NM) CAI_BANK_NM,\n"+
        "TRIM(A.CAI_BRANCH_NM) CAI_BRANCH_NM,\n"+
        "TRIM(A.CAI_ACC_TP) CAI_ACC_TP,\n"+
        "TRIM(A.CAI_ACC_NO) CAI_ACC_NO,\n"+
        "TRIM(A.CAI_ACC_NM) CAI_ACC_NM,\n"+
        "TRIM(A.PAY_MTD) PAY_MTD,\n"+
        "TRIM(A.CUST_REMIT_MTD) CUST_REMIT_MTD,\n"+
        "TRIM(A.CUST_BANK_CHARGE_TP) CUST_BANK_CHARGE_TP,\n"+
        "TRIM(A.CUST_CMPNY_BANK_CHARGE_EXPL) CUST_CMPNY_BANK_CHARGE_EXPL,\n"+
        "TRIM(A.CAI_ACC_REMARK) CAI_ACC_REMARK,\n"+
        "A.INVOECE_DATE_PICC_TP INVOECE_DATE_PICC_TP,\n"+
        "TRIM(A.INVOICE_DATE_REMARK) INVOICE_DATE_REMARK,\n"+
        "A.CLOSE_DATE_PICC_TP CLOSE_DATE_PICC_TP,\n"+
        "TRIM(A.INCOME_CLOSE_DATE_REMARK) INCOME_CLOSE_DATE_REMARK,\n"+
        "A.INCOME_DATE_PICC_TP INCOME_DATE_PICC_TP,\n"+
        "TRIM(A.INCOME_DATE_REMARK) INCOME_DATE_REMARK,\n"+
        "TRIM(A.SITE_TP) SITE_TP,\n"+
        "TRIM(A.CRI_BANK_CD) CRI_BANK_CD,\n"+
        "TRIM(A.CRI_BRANCH_CD) CRI_BRANCH_CD,\n"+
        "TRIM(A.CRI_ACC_TP) CRI_ACC_TP,\n"+
        "TRIM(B.ACC_NO) ACC_NO,\n"+
        "TRIM(B.ACC_NM) ACC_NM,\n"+
        "TRIM(A.INCOME_MTD) INCOME_MTD,\n"+
        "TRIM(A.OUR_CMPNY_REMIT_MTD) OUR_CMPNY_REMIT_MTD,\n"+
        "TRIM(A.OUR_CMPNY_BANK_CHARGE_TP) OUR_CMPNY_BANK_CHARGE_TP,\n"+
        "TRIM(A.OUR_CMPNY_BANK_CHARGE_EXPL) OUR_CMPNY_BANK_CHARGE_EXPL,\n"+
        "TRIM(A.CRI_ACC_REMARK) CRI_ACC_REMARK\n"+
        "FROM CUST A, CUST_BAI B\n";
strBuff.append(SQL);
}
Date after = new Date();
b = after.getTime();
System.out.println(b-a);
Date before2 = new Date();
long a2,b2;
a2 = before2.getTime();
StringBuffer strBuff2 = new StringBuffer(1280000); for (int i2=0; i2<1000; i2++ )
{
        strBuff2.append("SELECT \n");
        strBuff2.append("TRIM(A.FML_CUST_NM) FML_CUST_NM,\n");
        strBuff2.append("TRIM(A.CAI_BANK_NM) CAI_BANK_NM,\n");
        strBuff2.append("TRIM(A.CAI_BRANCH_NM) CAI_BRANCH_NM,\n");
        strBuff2.append("TRIM(A.CAI_ACC_TP) CAI_ACC_TP,\n");
        strBuff2.append("TRIM(A.CAI_ACC_NO) CAI_ACC_NO,\n");
        strBuff2.append("TRIM(A.CAI_ACC_NM) CAI_ACC_NM,\n");
        strBuff2.append("TRIM(A.PAY_MTD) PAY_MTD,\n");
       strBuff2.append("TRIM(A.CUST_REMIT_MTD) CUST_REMIT_MTD,\n");
        strBuff2.append("TRIM(A.CUST_BANK_CHARGE_TP) CUST_BANK_CHARGE_TP,\n");
        strBuff2.append("TRIM(A.CUST_CMPNY_BANK_CHARGE_EXPL) CUST_CMPNY_BANK_CHARGE_EXPL,\n");
        strBuff2.append("TRIM(A.CAI_ACC_REMARK) CAI_ACC_REMARK,\n");
        strBuff2.append("A.INVOECE_DATE_PICC_TP INVOECE_DATE_PICC_TP,\n");
        strBuff2.append("TRIM(A.INVOICE_DATE_REMARK) INVOICE_DATE_REMARK,\n");
        strBuff2.append("A.CLOSE_DATE_PICC_TP CLOSE_DATE_PICC_TP,\n");
        strBuff2.append("TRIM(A.INCOME_CLOSE_DATE_REMARK) INCOME_CLOSE_DATE_REMARK,\n");
        strBuff2.append("A.INCOME_DATE_PICC_TP INCOME_DATE_PICC_TP,\n");
        strBuff2.append("TRIM(A.INCOME_DATE_REMARK) INCOME_DATE_REMARK,\n");
        strBuff2.append("TRIM(A.SITE_TP) SITE_TP,\n");
        strBuff2.append("TRIM(A.CRI_BANK_CD) CRI_BANK_CD,\n");
        strBuff2.append("TRIM(A.CRI_BRANCH_CD) CRI_BRANCH_CD,\n");
        strBuff2.append("TRIM(A.CRI_ACC_TP) CRI_ACC_TP,\n");
        strBuff2.append("TRIM(B.ACC_NO) ACC_NO,\n");
        strBuff2.append("TRIM(B.ACC_NM) ACC_NM,\n");
        strBuff2.append("TRIM(A.INCOME_MTD) INCOME_MTD,\n");
        strBuff2.append("TRIM(A.OUR_CMPNY_REMIT_MTD) OUR_CMPNY_REMIT_MTD,\n");
        strBuff2.append("TRIM(A.OUR_CMPNY_BANK_CHARGE_TP) OUR_CMPNY_BANK_CHARGE_TP,\n");
        strBuff2.append("TRIM(A.OUR_CMPNY_BANK_CHARGE_EXPL) OUR_CMPNY_BANK_CHARGE_EXPL,\n");
        strBuff2.append("TRIM(A.CRI_ACC_REMARK) CRI_ACC_REMARK\n");
        strBuff2.append("FROM CUST A, CUST_BAI B\n");
} Date after2 = new Date();
b2 = after2.getTime();
System.out.println(b2-a2); }
}

真奇怪，我又作了实验，发现从(1)到(4)越来越慢。
哪位高人解释一下(1)
String SQL = "";
SQL = "A" + "B";
(2)
String SQL = "A" + "B";
(3)
StringBuffer sbuff = new StringBuffer(容量)
StringBuff.append("A");
StringBuff.append("B");
(4)
StringBuffer sbuff = new StringBuffer( )
StringBuff.append("A");
StringBuff.append("B");

可能的原因是：常量字符串+常量字符串时，编译器将他们合并成为一个常量字符串，也就是说"cosnt a"+"const b"与"const aconst b"的效果是一样的。
但是象这样子也非常慢：
SQL = SQL + "const a";
SQL = SQL + "const b";
这时会生成中间临时的String对象。

//测试string+与stringbuffer.append，结论，后者的性能大于前者500倍以上
    String ls_test="",ls_temp="fuckck";
    long ll_start,ll_end;
    StringBuffer lsb_test=new Strinuffer();
    ll_start=System.currentTimeMill();
    for (int i=0;i<500000;i++)
        lsb_test.append(ls_temp);
    ll_end=System.currentTimeMillis;
    System.out.println("consume:"+(_end-ll_start));

    ll_start=System.currentTimeMill();
    for (int i=0;i<5000;i++)
        ls_test+=ls_temp;
    ll_end=System.currentTimeMillis;
    System.out.println("consume:"+(_end-ll_start));
    //System.out.println(lsb_test);

调试易

象这种用加号定义一个String变量是否效率很低，我的意思是这里会不会生成多个中间String变量

解决方案 »