这篇文章给大家介绍java中怎么利用正则表达式获取sql中的表名，内容非常详细，感兴趣的小伙伴们可以参考借鉴，希望对大家能有所帮助。
/**
* 表名解析
*/
private final static Pattern P = Pattern
.compile("\\s+from\\s+(\\w\\.)?(\\w+)(\\s+|,)(\\w*)|\\s+join\\s+(\\w\\.)?(\\w+)\\s+(\\w\\s)*on", Pattern.CASE_INSENSITIVE);
/**
* 预处理sql
* @param sql sql
* @return 处理后sql
* @throws Exception 异常
*/
private static String preHandleSql(String sql,boolean removeBrackets) throws Exception {
// 1. 统一换行符
if (sql.contains("\r\n")) {
sql = sql.replaceAll("\r\n", "\n");
} else {
sql = sql.replaceAll("\r", "\n");
}
// 2. 去掉注释
String[] rows = sql.split("\n");
StringBuilder sb = new StringBuilder(sql.length());
for (String row : rows) {
int indexOfComment = row.indexOf("--");
if (indexOfComment == -1) {
sb.append(row).append("\n");
} else {
sb.append(row, 0, indexOfComment).append("\n");
}
}
if (removeBrackets) {
sql = removeBrackets(sb);
}
// 4. 找到第一个select和from，其间的字段即是要查询的字段列表
return sql.toLowerCase();
}
/**
* 括号消除：消除括号和括号中包含的内容，非贪心模式
* 例如：
* 输入select (ssss(sadf)sdfsdf(sdssf(sssdf)dsssf)dsslf) as a, (asdfsdf) as b, (asdfsdf) as c from dsfdf
* a组 b组 a组 c组 d组 b组 c组 d组 m组 m组 x组 x组
* A组 A组 B组 B组 C组 C组
* 去掉select (去 掉 部 分) as a, (去掉部分) as b, (去掉部分) as c from dsfdf
* 输出select as a, as b, as c from dsfdf
*
* @param str 消除前的字符串
* @return 消除后的字符串
* @throws Exception 括号不匹配
*/
private static String removeBrackets(StringBuilder str) throws Exception {
// 1. 收集括号组
List<int[]> bracketsContainer = new ArrayList<>();
collectBrackets(str, 0, bracketsContainer);
if (bracketsContainer.isEmpty()) {
return str.toString();
}
// 2. 消除括号组
StringBuilder newStr = new StringBuilder(str.length());
int groupSize = bracketsContainer.size();
for (int i = 0; i < groupSize; i++) {
int[] currentBrackets = bracketsContainer.get(i);
if (i == 0) {
// 刚到第一组
newStr.append(str.subSequence(0, currentBrackets[0]));
}
if (i + 1 == groupSize) {
// 已到最后一组
newStr.append(str.subSequence(currentBrackets[1] + 1, str.length()));
} else {
// 未到最后一组
int[] nextBrackets = bracketsContainer.get(i + 1);
newStr.append(str.subSequence(currentBrackets[1] + 1, nextBrackets[0]));
}
}
return newStr.toString();
}
/**
* 收集括号组
*
* @param str 消除前的字符串
* @param fromIndex 从哪开始找括号
* @param bracketsContainer 括号组容器
* @throws Exception 括号不匹配、括号嵌套层级过多
*/
private static void collectBrackets(StringBuilder str, int fromIndex, List<int[]> bracketsContainer) throws Exception {
int firstLeftBracket = str.indexOf("(", fromIndex + 1);
int nextLeftBracket = firstLeftBracket;
// SQL中不包含左括号时，直接返回
if (firstLeftBracket == -1) {
return;
}
// 括号层级（因为前面已经找到一个左括号，所以初始值为1）
int level = 1;
int nextRightBracket = str.indexOf(")", fromIndex + 1);
if (nextRightBracket == -1) {
throw new Exception("括号不匹配");
}
// 避免死循环
int maxLevel = 1000;
do {
int tempLeftBracket = str.indexOf("(", nextLeftBracket + 1);
if (tempLeftBracket == -1 || tempLeftBracket > nextRightBracket) {
// 找不到下一个左括号或者下一个左括号已属于下一个括号组
break;
} else {
nextLeftBracket = tempLeftBracket;
}
nextRightBracket = str.indexOf(")", nextRightBracket + 1);
if (nextRightBracket == -1) {
throw new Exception("括号不匹配");
}
level++;
} while (level <= maxLevel);
if (level >= maxLevel) {
throw new Exception("括号嵌套层级过多");
}
// 把收集到的括号组放入容器
bracketsContainer.add(new int[] {firstLeftBracket, nextRightBracket});
// 递归
collectBrackets(str, nextRightBracket, bracketsContainer);
}
/**
* 解析sql中的表名
* @param sql sql
* @return 表名
* @throws Exception 异常
*/
public static List<String> parseSqlRefTables(String sql) throws Exception {
List<String> tableNames = Lists.newArrayList();
String newSql = StringUtils.replaceAll(preHandleSql(sql, false),"\n"," ");
Matcher m = p.matcher(newSql);
while (m.find()) {
tableNames.add(m.group());
}
List<String> result = Lists.newArrayList();
for (String tableName : tableNames) {
String trimTableName = StringUtils.split(tableName, " ")[1].trim();
if (trimTableName.contains(",")) {
result.addAll(Arrays.asList(StringUtils.split(trimTableName, ",")));
} else {
result.add(trimTableName);
}
}
return result.parallelStream().distinct().collect(Collectors.toList());
}
public static void main(String[] args) throws Exception {
String data1 = "SELECT dma_t.\"id\", dma_t.\"name\", dma_t.area_border, dma_t.lat_lng, dma_t.level, leak_t.lossf, leak_t.\"day\", leak_t.avgf, leak_t.background_loss, leak_t.nmf, leak_t.normal_use, leak_t.supply FROM la_leak_t leak_t, ( SELECT MAX (dma_id) AS dma_id, MAX (\"day\") AS \"day\" FROM la_leak_t WHERE org_id = ${orgId} GROUP BY dma_id ) recent_t, mdm_dmaarea_m_t dma_t WHERE leak_t.dma_id = recent_t.dma_id AND leak_t.\"day\" = recent_t.\"day\" AND leak_t.dma_id = dma_t.\"id\"";
List<String> tableNames = parseSqlRefTables(data1);
tableNames.forEach(System.out::println);
}关于java中怎么利用正则表达式获取sql中的表名就分享到这里了，希望以上内容可以对大家有一定的帮助，可以学到更多知识。如果觉得文章不错，可以把它分享出去让更多的人看到。
免责声明：本站发布的内容（图片、视频和文字）以原创、转载和分享为主，文章观点不代表本网站立场，如果涉及侵权请联系站长邮箱：is@yisu.com进行举报，并提供相关证据，一经查实，将立刻删除涉嫌侵权内容。
