如何在func.count上執行HAVING？

Question

這是我的代碼：

    charge_count = func.count(ChargeLog).label('charge_count')
    q = (session.query(PaymentProfile, charge_count)
         .outerjoin(ChargeLog, ChargeLog.profile_id == PaymentProfile._id)
         .filter(PaymentProfile.date_expiry_utc >= datetime.datetime.utcnow() - datetime.timedelta(days=(TOTAL_TIMES_TO_CHARGE - charging_attempt_0idx)))
         .filter(PaymentProfile.status == payments.ProfileStatus.ACTIVE)
         .filter(ChargeLog.date_created_utc > PaymentProfile.date_charged_utc)
         .group_by(PaymentProfile)
         .having(charge_count == charging_attempt_0idx))

但是，我收到sqlalchemy.exc.ProgrammingError: (psycopg2.ProgrammingError) can't adapt type 'DeclarativeMeta'的錯誤sqlalchemy.exc.ProgrammingError: (psycopg2.ProgrammingError) can't adapt type 'DeclarativeMeta' 。

特別：

sqlalchemy.exc.ProgrammingError：（psycopg2.ProgrammingError）無法適應類型'DeclarativeMeta'[SQL：'SELECT payment_profile._id AS payment_profile__id，payment_profile.sb_org_id AS payment_profile_sb_org_id，payment_profile.plan_id AS payment_profile_plan_us，payment_profile.stats date_created_utc AS payment_profile_date_created_utc，payment_profile.date_updated_utc AS payment_profile_date_updated_utc，payment_profile.date_charged_utc AS payment_profile_date_charged_utc，payment_profile.date_expiry_utc AS payment_profile_date_expiry_utc，payment_profile.payment_method_id AS payment_profile_payment_method_id，計數（％（PARAM_1）S）AS charge_count \\ n從payment_profile LEFT OUTER JOIN charge_log ON charge_log.profile_id = Payment_profile._id \\ nWHEREpayment_profile.date_expiry_utc> =％（date_expiry_utc_1）s AND payment_profile.status =％（status_1）s AND charge_log.date_created_utc>付款文件.date_charged_utc GROUP BY payment_profile._id，pa yment_profile.sb_org_id，payment_profile.plan_id，payment_profile.status，payment_profile.date_created_utc，payment_profile.date_updated_utc，payment_profile.date_charged_utc，payment_profile.date_expiry_utc，payment_profile.payment_method_id \\ nHAVING計數（％）（％）參數：{'date_expiry_utc_1'：datetime.datetime（2017，2，14，14，9，2，11，353027），'param_1'：，'status_1'：'active'，'param_2'：0}]]

我把它歸功於'param_1': <class 'sb_base.model.ChargeLog'>通過該charge_count不翻譯成標簽在被解析having條件。

我怎樣才能解決這個問題？

Answer 1

您需要使用ChargeLog表中的一列，而不是表映射對象本身。 使用主鍵：

# assumption: ChargeLog.id is the primary key
charge_count = func.count(ChargeLog.id).label('charge_count')

這將在SELECT列和HAVING子句中插入相同的表達式； 您通常不能在HAVING使用標簽：

SELECT
  -- more columns
  count(charge_log.id) AS charge_count
-- FROM, OUTER JOIN, WHERE, GROUP_BY
HAVING count(charge_log.id)