hotime/db/identifier.go

package db

import (
	"regexp"
	"strings"
)

// IdentifierProcessor 标识符处理器
// 用于处理表名、字段名的前缀添加和引号转换
type IdentifierProcessor struct {
	dialect Dialect
	prefix  string
}

// NewIdentifierProcessor 创建标识符处理器
func NewIdentifierProcessor(dialect Dialect, prefix string) *IdentifierProcessor {
	return &IdentifierProcessor{
		dialect: dialect,
		prefix:  prefix,
	}
}

// 系统数据库列表，这些数据库不添加前缀
var systemDatabases = map[string]bool{
	"INFORMATION_SCHEMA": true,
	"information_schema": true,
	"mysql":              true,
	"performance_schema": true,
	"sys":                true,
	"pg_catalog":         true,
	"pg_toast":           true,
}

// ProcessTableName 处理表名（添加前缀+引号）
// 输入: "order" 或 "`order`" 或 "\"order\"" 或 "INFORMATION_SCHEMA.TABLES"
// 输出: "`app_order`" (MySQL) 或 "\"app_order\"" (PostgreSQL/SQLite)
// 对于 database.table 格式，会分别处理，系统数据库不添加前缀
func (p *IdentifierProcessor) ProcessTableName(name string) string {
	// 去除已有的引号
	name = p.stripQuotes(name)

	// 检查是否包含空格（别名情况，如 "order AS o"）
	if strings.Contains(name, " ") {
		// 处理别名情况
		parts := strings.SplitN(name, " ", 2)
		tableName := p.stripQuotes(parts[0])
		alias := parts[1]
		// 递归处理表名部分（可能包含点号）
		return p.ProcessTableName(tableName) + " " + alias
	}

	// 检查是否包含点号（database.table 格式）
	if strings.Contains(name, ".") {
		parts := p.splitTableColumn(name)
		if len(parts) == 2 {
			dbName := p.stripQuotes(parts[0])
			tableName := p.stripQuotes(parts[1])
			// 系统数据库不添加前缀
			if systemDatabases[dbName] {
				return p.dialect.QuoteIdentifier(dbName) + "." + p.dialect.QuoteIdentifier(tableName)
			}
			// 非系统数据库，只给表名添加前缀
			return p.dialect.QuoteIdentifier(dbName) + "." + p.dialect.QuoteIdentifier(p.prefix+tableName)
		}
	}

	// 添加前缀和引号
	return p.dialect.QuoteIdentifier(p.prefix + name)
}

// ProcessTableNameNoPrefix 处理表名（只添加引号，不添加前缀）
// 用于已经包含前缀的情况
func (p *IdentifierProcessor) ProcessTableNameNoPrefix(name string) string {
	name = p.stripQuotes(name)
	if strings.Contains(name, " ") {
		parts := strings.SplitN(name, " ", 2)
		tableName := p.stripQuotes(parts[0])
		alias := parts[1]
		return p.dialect.QuoteIdentifier(tableName) + " " + alias
	}
	return p.dialect.QuoteIdentifier(name)
}

// ProcessColumn 处理 table.column 格式
// 输入: "name" 或 "order.name" 或 "`order`.name" 或 "`order`.`name`"
// 输出: "`name`" 或 "`app_order`.`name`" (MySQL) / "\"app_order\".\"name\"" (DM/PG)
func (p *IdentifierProcessor) ProcessColumn(name string) string {
	if !strings.Contains(name, ".") {
		return p.dialect.QuoteIdentifier(p.stripQuotes(name))
	}

	parts := p.splitTableColumn(name)
	if len(parts) == 2 {
		tableName := p.stripQuotes(parts[0])
		columnName := p.stripQuotes(parts[1])
		return p.dialect.QuoteIdentifier(p.prefix+tableName) + "." + p.dialect.QuoteIdentifier(columnName)
	}

	return p.convertQuotes(name)
}

// ProcessColumnNoPrefix 处理 table.column 格式（不添加前缀）
func (p *IdentifierProcessor) ProcessColumnNoPrefix(name string) string {
	if !strings.Contains(name, ".") {
		return p.dialect.QuoteIdentifier(p.stripQuotes(name))
	}

	parts := p.splitTableColumn(name)
	if len(parts) == 2 {
		tableName := p.stripQuotes(parts[0])
		columnName := p.stripQuotes(parts[1])
		return p.dialect.QuoteIdentifier(tableName) + "." + p.dialect.QuoteIdentifier(columnName)
	}

	return p.convertQuotes(name)
}

// ProcessConditionString 智能解析条件字符串（如 ON 条件）
// 输入: "user.id = order.user_id AND order.status = 1"
// 输出: "`app_user`.`id` = `app_order`.`user_id` AND `app_order`.`status` = 1" (MySQL)
// 输出: "\"app_user\".\"id\" = \"app_order\".\"user_id\" AND \"app_order\".\"status\" = 1" (DM/PG)
func (p *IdentifierProcessor) ProcessConditionString(condition string) string {
	if condition == "" {
		return condition
	}

	result := condition

	// 首先处理已有完整引号的情况 `table`.`column` 或 "table"."column"
	fullyQuotedPattern := regexp.MustCompile("[`\"]([a-zA-Z_][a-zA-Z0-9_]*)[`\"]\\.[`\"]([a-zA-Z_][a-zA-Z0-9_]*)[`\"]")
	result = fullyQuotedPattern.ReplaceAllStringFunc(result, func(match string) string {
		parts := fullyQuotedPattern.FindStringSubmatch(match)
		if len(parts) == 3 {
			tableName := parts[1]
			colName := parts[2]
			return p.dialect.QuoteIdentifier(p.prefix+tableName) + "." + p.dialect.QuoteIdentifier(colName)
		}
		return match
	})

	// 然后处理部分引号的情况 `table`.column 或 "table".column
	quotedTablePattern := regexp.MustCompile("[`\"]([a-zA-Z_][a-zA-Z0-9_]*)[`\"]\\.([a-zA-Z_][a-zA-Z0-9_]*)(?:[^`\"]|$)")
	result = quotedTablePattern.ReplaceAllStringFunc(result, func(match string) string {
		parts := quotedTablePattern.FindStringSubmatch(match)
		if len(parts) >= 3 {
			tableName := parts[1]
			colName := parts[2]
			suffix := ""
			if len(match) > len(parts[0])-1 {
				lastChar := match[len(match)-1]
				if lastChar != '`' && lastChar != '"' && !isIdentChar(lastChar) {
					suffix = string(lastChar)
				}
			}
			return p.dialect.QuoteIdentifier(p.prefix+tableName) + "." + p.dialect.QuoteIdentifier(colName) + suffix
		}
		return match
	})

	// 最后处理无引号的情况 table.column
	unquotedPattern := regexp.MustCompile(`([^` + "`" + `"\w]|^)([a-zA-Z_][a-zA-Z0-9_]*)\.([a-zA-Z_][a-zA-Z0-9_]*)([^` + "`" + `"\w(]|$)`)
	result = unquotedPattern.ReplaceAllStringFunc(result, func(match string) string {
		parts := unquotedPattern.FindStringSubmatch(match)
		if len(parts) >= 5 {
			prefix := parts[1]
			tableName := parts[2]
			colName := parts[3]
			suffix := parts[4]
			return prefix + p.dialect.QuoteIdentifier(p.prefix+tableName) + "." + p.dialect.QuoteIdentifier(colName) + suffix
		}
		return match
	})

	return result
}

// isIdentChar 判断是否是标识符字符
func isIdentChar(c byte) bool {
	return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || (c >= '0' && c <= '9') || c == '_'
}

// ProcessFieldList 处理字段列表字符串
// 输入: "order.id, user.name AS uname, COUNT(*)"
// 输出: "`app_order`.`id`, `app_user`.`name` AS uname, COUNT(*)" (MySQL)
func (p *IdentifierProcessor) ProcessFieldList(fields string) string {
	if fields == "" || fields == "*" {
		return fields
	}

	// 使用与 ProcessConditionString 相同的逻辑
	return p.ProcessConditionString(fields)
}

// stripQuotes 去除标识符两端的引号（反引号或双引号）
func (p *IdentifierProcessor) stripQuotes(name string) string {
	name = strings.TrimSpace(name)
	// 去除反引号
	if strings.HasPrefix(name, "`") && strings.HasSuffix(name, "`") {
		return name[1 : len(name)-1]
	}
	// 去除双引号
	if strings.HasPrefix(name, "\"") && strings.HasSuffix(name, "\"") {
		return name[1 : len(name)-1]
	}
	return name
}

// splitTableColumn 分割 table.column 格式
// 支持: table.column, `table`.column, `table`.`column`, "table".column 等
func (p *IdentifierProcessor) splitTableColumn(name string) []string {
	// 先尝试按点号分割
	dotIndex := -1

	// 查找不在引号内的点号
	inQuote := false
	quoteChar := byte(0)
	for i := 0; i < len(name); i++ {
		c := name[i]
		if c == '`' || c == '"' {
			if !inQuote {
				inQuote = true
				quoteChar = c
			} else if c == quoteChar {
				inQuote = false
			}
		} else if c == '.' && !inQuote {
			dotIndex = i
			break
		}
	}

	if dotIndex == -1 {
		return []string{name}
	}

	return []string{name[:dotIndex], name[dotIndex+1:]}
}

// convertQuotes 将已有的引号转换为当前方言的引号格式
func (p *IdentifierProcessor) convertQuotes(name string) string {
	quoteChar := p.dialect.QuoteChar()
	// 替换反引号
	name = strings.ReplaceAll(name, "`", quoteChar)
	// 如果目标是反引号，需要替换双引号
	if quoteChar == "`" {
		name = strings.ReplaceAll(name, "\"", quoteChar)
	}
	return name
}

// GetDialect 获取方言
func (p *IdentifierProcessor) GetDialect() Dialect {
	return p.dialect
}

// GetPrefix 获取前缀
func (p *IdentifierProcessor) GetPrefix() string {
	return p.prefix
}