查詢sql如下：

 select 
(select to_char(wm_concat(a.description)) from project_report_detail a 
 where project_report_id = t.id) as description2 ,
 t.* 
 from project_report t  
 where 1=1  order by t.dept_id 

然後由於project_report_detail表的description欄位是多條資料拼接成一個欄位，拼接出來的結果字串超過了sql的varchar2型別4000字元，所以報了此錯：

緩衝區對於 CLOB 到 CHAR 轉換或 BLOB 到 RAW 轉換而言太小 (實際: 4763, 最大: 4000)

百度了一下找到了靠譜的解決辦法，借鑑部落格：http://www.linuxboy.net/linuxjc/137415.html

我的解決辦法如下：查詢的sql不再用to_char轉成字串，直接查出拼接欄位的clob型別

 select 
(select wm_concat(a.description) from project_report_detail a 
 where project_report_id = t.id) as description2 ,
 t.* 
 from project_report t  
 where 1=1  order by t.dept_id 

　　然後在java中，用上面部落格的方法處理成String型別：

private String ClobtoString(Clob clob){
        String reString = "";
        Reader is = null;
        try {
            is = clob.getCharacterStream();
        } catch (Exception e) {
            e.printStackTrace();
        }
        // 得到流
        BufferedReader br = new
 
 BufferedReader(is);
        String s = null;
        try {
            s = br.readLine();
        } catch (Exception e) {
            e.printStackTrace();
        }
        StringBuffer sb = new StringBuffer();
        while (s != null) {
            // 執行迴圈將字串全部取出付值給StringBuffer由StringBuffer轉成STRING
            sb.append(s);
            try {
                s = br.readLine();
            } catch (Exception e) {
                // TODO Auto-generated catch block
                e.printStackTrace();
            }
        }
        reString = sb.toString();
        return reString;
    }

成功解決，親測有效