大数据量写文件 Java heap space

Question

[code="java"]
Statement stmt = ds.getConnection().createStatement();
stmt.setFetchSize(2000);
ResultSet rs = stmt.executeQuery(sqlQuery);
pw = new PrintWriter(new FileWriter(fileW, true));
StringBuilder dataStr2 = new StringBuilder("");
int c=0;
for (; rs.next(); pw.print(dataStr2.append("\r\n").toString())) {
dataStr2 = new StringBuilder("");
for (int i = 1; i <= 8; i++)
dataStr2.append(rs.getString(i));
if (c++ % 2000 == 0)
pw.flush();
}
[/code]
请问数据量上百万的，以上代码怎么优化！

广告离我远点 -_#

Answer 1

大致修改了下，如果还是很慢，建议你研究下你的SQL，因为写文件是耗不了多少时间的

[code="java"]Statement stmt = ds.getConnection().createStatement();

stmt.setFetchSize(2000);

ResultSet rs = stmt.executeQuery(sqlQuery);

pw = new PrintWriter(new FileWriter(fileW, true));

StringBuilder dataStr2 = new StringBuilder("");

int c=0;

for (; rs.next(); ) {

dataStr2 = dataStr2.delete(0, dataStr2.length()); // 避免构造大量对象实例
for (int i = 1; i <= 8; i++)

dataStr2.append(rs.getString(i));

if (c++ % 2000 == 0) {
pw.flush();
// 200条数据写一次,没必要没读一条写一次
pw.print(dataStr2.append("\r\n").toString())) ;

}
}[/code]