QingLong/ClickHouse/mappers/data_mapper.py

from datetime import datetime
from utils.logger import configure_logger
logger = configure_logger()

class DataMapper:
    def __init__(self, mysql_conn, table):
        self.mysql_conn = mysql_conn
        self.table = table
        self.date_columns = []
        self.uint64_columns = []
        self._analyze_schema()
        self.min_date = datetime(1970, 1, 1)
        self.max_date = datetime(2105, 12, 31, 23, 59, 59)

    def _analyze_schema(self):
        """分析表结构自动识别字段类型"""
        schema_query = f"""
        SELECT COLUMN_NAME, DATA_TYPE, COLUMN_TYPE 
        FROM INFORMATION_SCHEMA.COLUMNS 
        WHERE TABLE_SCHEMA = DATABASE() 
          AND TABLE_NAME = '{self.table}'
        """

        with self.mysql_conn.connect().cursor() as cursor:
            cursor.execute(schema_query)
            for col_name, data_type, col_type in cursor.fetchall():
                # 识别日期时间字段
                if data_type in ('datetime', 'timestamp', 'date'):
                    self.date_columns.append(col_name)

                # 识别无符号整数字段（匹配BIGINT UNSIGNED等）
                if 'unsigned' in col_type.lower() and 'int' in data_type:
                    self.uint64_columns.append(col_name)


        #logger.info(f"自动识别字段类型 - 日期字段: {self.date_columns}")
        #logger.info(f"自动识别字段类型 - 无符号整数字段: {self.uint64_columns}")
    def map_row(self, columns, row):
        row_dict = dict(zip(columns, row))
        return {col: self._map_value(col, val) for col, val in row_dict.items()}

    def _map_value(self, col, value):
        if col in self.uint64_columns:
            return self._handle_uint64(value)
        elif col in self.date_columns:
            return self._handle_datetime(value)
        elif isinstance(value, str):
            return value.strip()
        return value

    def _handle_uint64(self, value):
        try:
            return int(float(value)) if value not in (None, '', 'NULL') else 0
        except:
            return 0

    def _handle_datetime(self, value):
        dt = self._parse_datetime(value)
        return dt if dt else self.min_date

    def _parse_datetime(self, value):
        if value in (None, 0, '0', '0.0', '0.00', '', 'null', 'NULL'):
            return self.min_date

        try:
            str_value = str(value).strip()
            for fmt in ('%Y-%m-%d %H:%M:%S', '%Y-%m-%d', '%Y%m%d%H%M%S', '%Y/%m/%d %H:%M:%S'):
                try:
                    parsed = datetime.strptime(str_value, fmt)
                    return self._clamp_datetime(parsed)
                except ValueError:
                    continue

            if str_value.isdigit():
                ts = int(str_value)
                if 1e12 < ts < 1e13:  # 毫秒级时间戳
                    parsed = datetime.fromtimestamp(ts / 1000)
                elif 1e9 < ts < 1e10:  # 秒级时间戳
                    parsed = datetime.fromtimestamp(ts)
                return self._clamp_datetime(parsed)

            return self.min_date
        except:
            return self.min_date

    def _clamp_datetime(self, dt):
        if dt < self.min_date:
            return self.min_date
        elif dt > self.max_date:
            return self.max_date
        return dt