Hi, I'm currently learning web2py for deployment on Google App Engine, and I've got a couple of questions about how the datastore works.
Let's say I've got a relational database with students, courses and enrolments. Enrolments is the join table that allows a many to many relationship between students and courses. For example: Students -------- id name Courses ------------ id title Enrolments ---------------- student_id course_id This allows me to show a list of courses for a student, and a list of users enrolled on a course, by using join queries. I can order these lists by the student name, and do pagination too (using LIMIT and OFFSET). However, as I understand it, in the Google datastore, you can't do joins like this. From what I've read you could do it with a list:reference field, which basically contains a list of IDs like this 1|2|3|4 etc., so the data would look like this: Student ----------- id name courses e.g. 1, "John Smith", 1|2|3 Course ---------- id title students e.g. 1, "Physics 101", 4|5|6 (and no enrolments table) Is that how data has to be modelled using the GAE datastore? If so, is it possible to do the following: - paginate the denormalised data, for example show a paginated list of students on a course? - have hundreds of IDs in the list:reference field (a course could have hundreds of students) - or is there an upper limit? Also, how are updates to these fields handled? For example, remove one student from a course that has hundreds enrolled? Are there mechanisms in web2py to handle this? Also, it would seem that the two list:reference fields would need to be kept synchronised somehow - if a student is deleted, any course that they were enrolled on needs to be updated too, which (with my mind still in relational database mode) would seem to be very inefficient. I think what I need is a high-level overview of how data needs to be designed for the GAE datastore ̣- however as I've found this forum to be very helpful I thought I'd ask my specific questions here first before I go off and watch some Google I/O videos. Any help in me trying to get my head around this is greatly appreciated! Many thanks Dave
