Question

class Journal(Base):
    __tablename__ = '_1SJOURN'
    rid = Column("ROW_ID", Integer, primary_key=True)
    journal_id = Column("IDJOURNAL", Integer)
    iddoc = Column("IDDOC", String)
    iddocdef = Column("IDDOCDEF", Integer) 
    datetime  = Column("DATE_TIME_IDDOC", String)

表的字段为DATE_TIME_IDDOC，如200809305DU3F4 65P，前8个字符为日期。我需要选择日期介于s = 20070503，e = 20090503之间的行。我试过了：

filter(
    and_(func.convert(sql.literal_column('DATE'), Journal.datetime, sql.literal_column('1,8')) > s,
         func.convert(sql.literal_column('DATE'), Journal.datetime, sql.literal_column('1,8')) < e))

Answer 1

首先，为什么不能将“datetime”列拆分为两列：datetime和“other_info”？您的表格结构未经过优化。如果它看起来如下所示，那么实现您想要的更容易。更容易和正确。

class Journal(Base):
    __tablename__ = '_1SJOURN'
    rid = Column("ROW_ID", Integer, primary_key=True)
    journal_id = Column("IDJOURNAL", Integer)
    iddoc = Column("IDDOC", String)
    iddocdef = Column("IDDOCDEF", Integer) 
    datetime = Column("DATE_TIME", DateTime)
    other = Column("OTHER", String)

# here I do not use `and_`, because it used by default
Session.query(Journal).filter(Journal.datetime > s, Journal.datetime < e)

但是，如果你真的想要它，那就是你的查询。原始SQL（对于Postgresql，其他DBMS具有相同的功能，但具有其他名称）：

SELECT * FROM journal WHERE to_timestamp(left(DATE_TIME_IDDOC, 8), 'YYYYMMDD') > '2008-10-01' AND to_timestamp(left(DATE_TIME_IDDOC, 8), 'YYYYMMDD') < '2010-01-20';

SQLAlchemy查询：

filter_col = func.to_timestamp(func.left(Journal.datetime, 8), 'YYYYMMDD')
Session.query(Journal).filter(filter_col > s, filter_col < e)

Answer 2

and_(func.convert(sql.literal_column('datetime'),func.substring(Journal.datetime, 1,8)) > s ,
             func.convert(sql.literal_column('datetime'), func.substring(Journal.datetime, 1, 8)) < e)).filter(

sqlalchemy如何投射混合场到目前为止？

2 个答案: