[BangPypers] SQLAlchemy and 'non-trivial' default values for a column

Vinod Kumar Narasimhaiah vinod.narasimhaiah at gmail.com
Wed Feb 6 06:45:39 CET 2013


I am not a python programmer, but how about this logic?

add new field called "consultation_count" to the same table

before every insert to the table:

Check if the date field on the last record (you might want to add a
timestamp field to get the last record easily) = today's date

If yes, then it's the same day-
    take the consultation count from the last record, increment it by one
and create the new record.
If No, then it's a start of the new day-
    set the counter to 1 and create the new record.



On Tue, Feb 5, 2013 at 9:58 PM, Sriram Karra <karra.etc at gmail.com> wrote:

> I have a Declarative table defined as follows:
>
> class Consultation(Base):
>     __tablename__ = 'consultation'
>
>     id         = Column(Integer, primary_key=True)
>     patient_id = Column(Integer, ForeignKey('patient.id'))
>     doctor_id  = Column(Integer, ForeignKey('doctor.id'))
>     date       = Column(Date(),  default=MyT.today())
>
> Each row in this table represents a single consultation instance of a
> patient seeing a doctor on a given day.
>
> I would like an additional attribute called "cid" that should be an
> auto-incrementing value representing how many-th consultation it was in
> that day. Basically it is an auto-incrementing counter, which gets reset to
> 0 at the start of a day (hence not unique, whereas the id will be unique).
> No row is ever deleted.
>
> How do I do achieve this with the least amount of additional database
> space? It is trivial to have another table with one column for date and
> another column for the total consultations thus far.
>
> Any help?
>
> -Karra
>
> P.S. This is for PRS - an open source patient record system for small
> clinics I am developing, and available at: https://github.com/skarra/PRS
> _______________________________________________
> BangPypers mailing list
> BangPypers at python.org
> http://mail.python.org/mailman/listinfo/bangpypers
>


More information about the BangPypers mailing list