手写一个简单的ElasticSearch SQL转换器(一)

   一.前言

   之前有个需求，是使ElasticSearch支持使用SQL进行简单查询，较新版本的ES已经支持该特性(不过貌似还是实验性质的？) ，而且git上也有elasticsearch-sql

插件，之所以决定手写一个，主要有两点原因：

      1. 目前用的ES版本较老

      2. elasticsearch-sql虽好，但比较复杂，代码也不易维护

      3. 练练手

 二.技术选型

   目前主流软件中通常使用ANTLR做词法语法分析，诸如著名的Hibernate，Spark，Hive等项目，之前因为工作原因也有所接触，不过如果只是解析标准SQL的话，

 其实还有更好的选择，如使用Hibernate或阿里巴巴的数据库Druid(Druid采用了手写词法语法分析器的方案，这种方式当然比自动ANTLR生成的解析器性能高得多), 这里

 我选择了第二种方案。

     开始之前先看下我们可以通过Druid拿到的SQL语言的抽象语法树：

                                                  图片：https://www.jianshu.com/p/437aa22ea3ca

 三.技术实现

     首先我们创建一个SqlParser类，主流程都在parse方法中，该方法负责将一个SQL字符串解析(顺便说一句，Druid支持多种SQL方言，这里我选择了MySQL)，

 并返回SearchSourceBuilder对象，这是一个ElasticSearch提供的DSL构建器，以该对象作为参数，ES client端即可发起对ES 服务端搜索请求。

/**
*
* @author fred
*
*/
public class SqlParser {
private final static String dbType = JdbcConstants.MYSQL;
private final static Logger logger = LoggerFactory.getLogger(SqlParser.class);
private SearchSourceBuilder builder;

public SqlParser(SearchSourceBuilder builder) {
this.builder = builder;
}
/**
* 将SQL解析为ES查询
*/
public SearchSourceBuilder parse(String sql) throws Exception {
if (Objects.isNull(sql)) {
throw new IllegalArgumentException("输入语句不得为空");
}
sql = sql.trim().toLowerCase();
List<SQLStatement> stmtList = SQLUtils.parseStatements(sql, dbType);
if (Objects.isNull(stmtList) || stmtList.size() != 1) {
throw new IllegalArgumentException("必须输入一句查询语句");
}
// 使用Parser解析生成AST
SQLStatement stmt = stmtList.get(0);
if (!(stmt instanceof SQLSelectStatement)) {
throw new IllegalArgumentException("输入语句须为Select语句");
}
SQLSelectStatement sqlSelectStatement = (SQLSelectStatement) stmt;
SQLSelectQuery sqlSelectQuery = sqlSelectStatement.getSelect().getQuery();
SQLSelectQueryBlock sqlSelectQueryBlock = (SQLSelectQueryBlock) sqlSelectQuery;

SQLExpr whereExpr = sqlSelectQueryBlock.getWhere();

// 生成ES查询条件
BoolQueryBuilder bridge = QueryBuilders.boolQuery();
bridge.must();

QueryBuilder whereBuilder = whereHelper(whereExpr); // 处理where
bridge.must(whereBuilder);
SQLOrderBy orderByExpr = sqlSelectQueryBlock.getOrderBy(); // 处理order by
if (Objects.nonNull(orderByExpr)) {
orderByHelper(orderByExpr, bridge);
}
builder.query(bridge);
return builder;
}

    主流程很简单，拿到SQL字符串后，直接通过Druid API将其转换为抽象语法树，我们要求输入语句必须为Select语句。接下来是对where语句和order by语句的处理，

  目前的难点其实主要在于如何将where语句映射到ES查询中。

     先从简单的看起，如何处理order by呢？SQL语句中 order by显然可以允许用户根据多字段排序，所以排序字段肯定是一个List<排序字段>，我们要做的就是将这个List映射到

SearchSourceBuilder对象中。见下面代码：

/**
* 处理所有order by字段
*
4      * @param orderByExpr
*/
private void orderByHelper(SQLOrderBy orderByExpr, BoolQueryBuilder bridge) {
List<SQLSelectOrderByItem> orderByList = orderByExpr.getItems(); // 待排序的列
for (SQLSelectOrderByItem sqlSelectOrderByItem : orderByList) {
if (sqlSelectOrderByItem.getType() == null) {
sqlSelectOrderByItem.setType(SQLOrderingSpecification.ASC); // 默认升序
}
12             String orderByColumn = sqlSelectOrderByItem.getExpr().toString();
builder.sort(orderByColumn,
sqlSelectOrderByItem.getType().equals(SQLOrderingSpecification.ASC) ? SortOrder.ASC
: SortOrder.DESC);
}
}

   通过Druid的API，我们很容易拿到了SQL语句中所有的排序字段，我们逐个遍历这些字段，拿到排序的列名字面量和顺序，传递给SearchSourceBuilder的sort方法，需注意的

是， 如果原始SQL中没有指定字段是顺序，我们默认升序。

    接下来我们处理稍微有点麻烦的where语句，因为SQL语句被解析成了语法树，很自然的我们想到使用递归方式进行处理。 而通常在处理递归问题的时候，

  我习惯于从递归的base case开始考虑，where语句中的运算符根据Druid API中的定义主要分为以下三种：

    1. 简单二元运算符：包括逻辑处理，如and, or 和大部分关系运算(后续会详细讲)

    2. between或not between运算符：我们可以简单的将其映射成ES中的Range Query

    3. in, not in 运算符: 可以简单的映射成ES中的Term Query

   通过Druid，我们可以很方便的获取每种运算中的运算符与操作数

/**
* 递归遍历“where”子树
*
* @return
*/
private QueryBuilder whereHelper(SQLExpr expr) throws Exception {
if (Objects.isNull(expr)) {
throw new NullPointerException("节点不能为空!");
}
BoolQueryBuilder bridge = QueryBuilders.boolQuery();
if (expr instanceof SQLBinaryOpExpr) { // 二元运算
SQLBinaryOperator operator = ((SQLBinaryOpExpr) expr).getOperator(); // 获取运算符
if (operator.isLogical()) { // and,or,xor
return handleLogicalExpr(expr);
} else if (operator.isRelational()) { // 具体的运算,位于叶子节点
return handleRelationalExpr(expr);
}
} else if (expr instanceof SQLBetweenExpr) { // between运算
SQLBetweenExpr between = ((SQLBetweenExpr) expr);
boolean isNotBetween = between.isNot(); // between or not between ?
String testExpr = between.testExpr.toString();
String fromStr = formatSQLValue(between.beginExpr.toString());
String toStr = formatSQLValue(between.endExpr.toString());
if (isNotBetween) {
bridge.must(QueryBuilders.rangeQuery(testExpr).lt(fromStr).gt(toStr));
} else {
bridge.must(QueryBuilders.rangeQuery(testExpr).gte(fromStr).lte(toStr));
}
return bridge;
} else if (expr instanceof SQLInListExpr) { // SQL的 in语句，ES中对应的是terms
SQLInListExpr siExpr = (SQL
576
InListExpr) expr;
boolean isNotIn = siExpr.isNot(); // in or not in?
String leftSide = siExpr.getExpr().toString();
List<SQLExpr> inSQLList = siExpr.getTargetList();
List<String> inList = new ArrayList<>();
for (SQLExpr in : inSQLList) {
String str = formatSQLValue(in.toString());
inList.add(str);
}
if (isNotIn) {
bridge.mustNot(QueryBuilders.termsQuery(leftSide, inList));
} else {
bridge.must(QueryBuilders.termsQuery(leftSide, inList));
}
return bridge;
}
return bridge;
}

   上述第一种情况比较复杂，首先我们先看看运算符是逻辑运算的情况：

    如下面的代码所示，如果运算符是逻辑运算符，我们需要对左右操作数分别递归，然后根据运算符类型归并结果：or可以映射成ES 中的Should，而and则映射成Must.

/**
* 逻辑运算符，目前支持and,or
*
* @return
* @throws Exception
*/
private QueryBuilder handleLogicalExpr(SQLExpr expr) throws Exception {
BoolQueryBuilder bridge = QueryBuilders.boolQuery();
SQLBinaryOperator operator = ((SQLBinaryOpExpr) expr).getOperator(); // 获取运算符
SQLExpr leftExpr = ((SQLBinaryOpExpr) expr).getLeft();
SQLExpr rightExpr = ((SQLBinaryOpExpr) expr).getRight();

// 分别递归左右子树，再根据逻辑运算符将结果归并
QueryBuilder leftBridge = whereHelper(leftExpr);
QueryBuilder rightBridge = whereHelper(rightExpr);
if (operator.equals(SQLBinaryOperator.BooleanAnd)) {
bridge.must(leftBridge).must(rightBridge);
} else if (operator.equals(SQLBinaryOperator.BooleanOr)) {
bridge.should(leftBridge).should(rightBridge);
}
return bridge;
}

   下面来讨论下第一种情况中，如果运算符是关系运算符的情况，我们知道，SQL中的关系运算主要就是一些比较运算符，诸如大于，小于，等于，Like等，这里我还加上了

正则搜索(不过貌似性能比较差，ES对正则搜索的限制颇多，不太建议使用)。

/**
* 大于小于等于正则
*
* @param expr
* @return
*/
private QueryBuilder handleRelationalExpr(SQLExpr expr) {
SQLExpr leftExpr = ((SQLBinaryOpExpr) expr).getLeft();
if (Objects.isNull(leftExpr)) {
throw new NullPointerException("表达式左侧不得为空");
}
String leftExprStr = leftExpr.toString();
String rightExprStr = formatSQLValue(((SQLBinaryOpExpr) expr).getRight().toString()); // TODO:表达式右侧可以后续支持方法调用
SQLBinaryOperator operator = ((SQLBinaryOpExpr) expr).getOperator(); // 获取运算符
QueryBuilder queryBuilder;
switch (operator) {
case GreaterThanOrEqual:
queryBuilder = QueryBuilders.rangeQuery(leftExprStr).gte(rightExprStr);
break;
case LessThanOrEqual:
queryBuilder = QueryBuilders.rangeQuery(leftExprStr).lte(rightExprStr);
break;
case Equality:
queryBuilder = QueryBuilders.boolQuery();
TermQueryBuilder eqCond = QueryBuilders.termQuery(leftExprStr, rightExprStr);
((BoolQueryBuilder) queryBuilder).must(eqCond);
break;
case GreaterThan:
queryBuilder = QueryBuilders.rangeQuery(leftExprStr).gt(rightExprStr);
break;
case LessThan:
queryBuilder = QueryBuilders.rangeQuery(leftExprStr).lt(rightExprStr);
break;
case NotEqual:
queryBuilder = QueryBuilders.boolQuery();
TermQueryBuilder notEqCond = QueryBuilders.termQuery(leftExprStr, rightExprStr);
((BoolQueryBuilder) queryBuilder).mustNot(notEqCond);
break;
case RegExp: // 对应到ES中的正则查询
queryBuilder = QueryBuilders.boolQuery();
RegexpQueryBuilder regCond = QueryBuilders.regexpQuery(leftExprStr, rightExprStr);
((BoolQueryBuilder) queryBuilder).mustNot(regCond);
break;
case NotRegExp:
queryBuilder = QueryBuilders.boolQuery();
RegexpQueryBuilder notRegCond = QueryBuilders.regexpQuery(leftExprStr, rightExprStr);
((BoolQueryBuilder) queryBuilder).mustNot(notRegCond);
break;
case Like:
queryBuilder = QueryBuilders.boolQuery();
MatchPhraseQueryBuilder likeCond = QueryBuilders.matchPhraseQuery(leftExprStr,
rightExprStr.replace("%", ""));
((BoolQueryBuilder) queryBuilder).must(likeCond);
break;
case NotLike:
queryBuilder = QueryBuilders.boolQuery();
MatchPhraseQueryBuilder notLikeCond = QueryBuilders.matchPhraseQuery(leftExprStr,
rightExprStr.replace("%", ""));
((BoolQueryBuilder) queryBuilder).mustNot(notLikeCond);
break;
default:
throw new IllegalArgumentException("暂不支持该运算符!" + operator.toString());
}
return queryBuilder;
}

    到这里我们就完成了SQL转ES DSL的功能了(其实只是简单查询的转换)，下面我们写几个Junit测试一下吧：

    首先是简单的比较运算：

public void normalSQLTest() {
String sql = "select * from test where time>= 1";
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
try {
searchSourceBuilder = new SqlParser(searchSourceBuilder).parse(sql);
} catch (Exception e) {
e.printStackTrace();
}
System.out.println(searchSourceBuilder);
SearchSourceBuilder builderToCompare = new SearchSourceBuilder();
QueryBuilder whereBuilder = QueryBuilders.rangeQuery("time").gte("1");
BoolQueryBuilder briage = QueryBuilders.boolQuery();
briage.must();
briage.must(whereBuilder);
builderToCompare.query(briage);
assertEquals(searchSourceBuilder,builderToCompare);
}

  下面是输出的ES 查询语句：

{
"query" : {
"bool" : {
"must" : [
{
"range" : {
"time" : {
"from" : "1",
"to" : null,
"include_lower" : true,
"include_upper" : true,
"boost" : 1.0
}
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
}
}

  再来个带排序的：

@Test
public void normalSQLWithOrderByTest() {
String sql = "select * from test where time>= 1 order by time desc";
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
try {
searchSourceBuilder = new SqlParser(searchSourceBuilder).parse(sql);
} catch (Exception e) {
e.printStackTrace();
}
System.out.println(searchSourceBuilder);
SearchSourceBuilder builderToCompare = new SearchSourceBuilder();
QueryBuilder whereBuilder = QueryBuilders.rangeQuery("time").gte("1");
BoolQueryBuilder briage = QueryBuilders.boolQuery();
briage.must();
briage.must(whereBuilder);
builderToCompare.sort("time",SortOrder.DESC);
builderToCompare.query(briage);
assertEquals(searchSourceBuilder,builderToCompare);
}

   between, in这些没什么区别，就不贴代码了，最后看看稍微复杂点儿，带逻辑运算的查询：

@Test
public void sqlLogicTest() {
String sql = "select * from test where raw_log not like"+"'%aaa' && b=1 or c=0";
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
try {
searchSourceBuilder = new SqlParser(searchSourceBuilder).parse(sql);
} catch (Exception e) {
e.printStackTrace();
}
System.out.println(searchSourceBuilder);
SearchSourceBuilder builderToCompare = new SearchSourceBuilder();
QueryBuilder builder =QueryBuilders.matchPhraseQuery("raw_log","aaa");

BoolQueryBuilder briage1 = QueryBuilders.boolQuery();//raw log not like
briage1.mustNot(builder);

BoolQueryBuilder briage2 = QueryBuilders.boolQuery();  //b=1
briage2.must(QueryBuilders.termQuery("b","1"));

BoolQueryBuilder briage3 = QueryBuilders.boolQuery();   // not like and b=1
briage3.must(briage1).must(briage2);

BoolQueryBuilder briage4 = QueryBuilders.boolQuery();    //c =0
briage4.must(QueryBuilders.termQuery("c","0"));

BoolQueryBuilder briage5 = QueryBuilders.boolQuery();  // not like and b =1 or c =0
briage5.should(briage3).should(briage4);

BoolQueryBuilder briage6 = QueryBuilders.boolQuery();
briage6.must();
briage6.must(briage5);
builderToCompare.query(briage6);
assertEquals(searchSourceBuilder,builderToCompare);
}

 下面是生成的查询语句：

{
"query" : {
"bool" : {
"must" : [
{
"bool" : {
"should" : [
{
"bool" : {
"must" : [
{
"bool" : {
"must_not" : [
{
"match_phrase" : {
"raw_log" : {
"query" : "aaa",
"slop" : 0,
"boost" : 1.0
}
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
},
{
"bool" : {
"must" : [
{
"term" : {
"b" : {
"value" : "1",
"boost" : 1.0
}
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
},
{
"bool" : {
"must" : [
{
"term" : {
"c" : {
"value" : "0",
"boost" : 1.0
}
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
}
],
"disable_coord" : false,
"adjust_pure_negative" : true,
"boost" : 1.0
}
}
}

   四.总结

     本篇文章主要讲述了如何使用Druid实现SQL语句转换ES DSL进行搜索的功能，后续文章中会陆续完善这个功能，实现诸如聚合查询，分页查询等功能。