Here is model:
class User(Base):
...
birthday = Column(Date, index=True) #in database it's like '1987-01-17'
...
I want to filter between two dates, for example to choose all users in interval 18-30 years.
How to implement it with SQLAlchemy?
I think of:
query = DBSession.query(User).filter(
and_(User.birthday >= '1988-01-17', User.birthday <= '1985-01-17')
)
# means age >= 24 and age <= 27
I know this is not correct, but how to do correct?
In fact, your query is right except for the typo: your filter is excluding all records: you should change the <= for >= and vice versa:
qry = DBSession.query(User).filter(
and_(User.birthday <= '1988-01-17', User.birthday >= '1985-01-17'))
# or same:
qry = DBSession.query(User).filter(User.birthday <= '1988-01-17').\
filter(User.birthday >= '1985-01-17')
Also you can use between:
qry = DBSession.query(User).filter(User.birthday.between('1985-01-17', '1988-01-17'))
'1985-01-17', you can also use datetime.date(1985, 1, 17) - may be easier to get at or work with in some environments.
BETWEEN operator, so the answer will be RDBMS dependent. but i do believe it is inclusive on all all modern RDBMSif you want to get the whole period:
from sqlalchemy import and_, func
query = DBSession.query(User).filter(and_(func.date(User.birthday) >= '1985-01-17'),\
func.date(User.birthday) <= '1988-01-17'))
That means range: 1985-01-17 00:00 - 1988-01-17 23:59
func.date does CAST on the column which removes the time from equation => this does NOT mean range with the time! This only works when time is NOT in the column - you must CAST it to Date like this, or make the column Date, once it's DateTime or timestamp - it's usually finished with 00:00 (both MySQL & PostgreSQL do this). More generic solution is not to cast, but to set the date you are sending to it's .endOfDay() so you actually send 1988-01-17 23:59:59 to the database compare :)from app import SQLAlchemyDB as db
Chance.query.filter(Chance.repo_id==repo_id,
Chance.status=="1",
db.func.date(Chance.apply_time)<=end,
db.func.date(Chance.apply_time)>=start).count()
it is equal to:
select
count(id)
from
Chance
where
repo_id=:repo_id
and status='1'
and date(apple_time) <= end
and date(apple_time) >= start
wish can help you.
Here is a version which works with the latest Flask version and uses flask-marshmallow.
from datetime import date, timedelta
from flask import jsonify
from app import db, ma
from app.models import User
from . import main
class UserSchema(ma.Schema):
class Meta:
fields = ('forename', 'surname', 'birthday', ...)
@main.route('/', methods=('GET',))
def get_users():
start_range = date.today() + timedelta(years=-30)
end_range = date.today() + timedelta(years=-18)
users = db.session.query(User).filter(User.birthday.between(start_range, end_range)).all()
users_schema = UserSchema(many=True)
return jsonify(users_schema.dump(users))
date.today() + timedelta(years=-18) should be named end_range