Yes, I would like to Usai Id and not the host / ip. We know that multi-tentant came to automating a service saas in web2py, but as literature, and we lack a bit of a hurry to create one field in the table auth.user.hospital and popular with the same content of created_by, so we have a function in the controller
def testegrupo (): rows = db (db.t_medico.f_hospital == auth.user.hospital). select () return dict (rows = rows) we know this is a hack of the tenant for web2py but it works. But I will implement the changes proposed in this DISCUSSION for the tenant to operate as a service saas. Ovidio Marinho Falcao Neto Web Developer [email protected] [email protected] ITJP - itjp.net.br 83 8826 9088 - Oi 83 9334 0266 - Claro Brasil 2012/6/29 Michele Comitini <[email protected]>: > Ovidio, > > I wonder if you could use auth.user.id or auth.membership.id as key > instead of host/ip? > > If the above is not viable, Massimo suggestion to have separate > databases is very good IMHO and you can build a Data Warehouse > http://en.wikipedia.org/wiki/Data_warehouse by using postgresql > clustering capabilities or one of many postgresql synchronization > tools. > > mic > > > 2012/6/29 Ovidio Marinho <[email protected]>: >> Well now after my DAL I have Instructions >> db = DAL ('postgres :/ / postgres: passwd @ localhost: 5432/mybd') >> db._common_fields=Field('request_tenant',default=request.env.http_host,writable=False,readable=False)] >> >> What happened? in all my tables I have request_tenant the field, and >> when I do a drive automatically writes the localhost: 8000. But now >> how is that several users with different hosts with dhcp changing its >> access each time with a different ip that can work? >> >> this post from Massimo does exactly what I want with the school he has >> provided for various and simultaneous access, I could spend the >> Massimo setting this app? >> >> http://www.mail-archive.com/[email protected]/msg64351.html >> >> Ovidio Marinho Falcao Neto >> Web Developer >> [email protected] >> [email protected] >> ITJP - itjp.net.br >> 83 8826 9088 - Oi >> 83 9334 0266 - Claro >> Brasil >> >> >> >> 2012/6/28 Massimo Di Pierro <[email protected]>: >>> Do me the issue is, should all data always be filtered by hospital. >>> If there should only be filtered in some cases and not in other, use >>> explicit queries >>> If some tables are always filtered and some are not, you may want to use the >>> tenant feature. >>> >>> >>> On Thursday, 28 June 2012 05:45:13 UTC-5, Ovidio Marinho wrote: >>>> >>>> I can not split into separate database, as the management reports need >>>> to filter all information. What I need is that the data are shown to >>>> hosptal (1) is only seen by him. No problem if all hospitals being in >>>> the same database, so is the question of security in view. When the >>>> login Hospital (1) is made only data related to hospital (1) should >>>> appear on the screen. >>>> >>>> >>>> >>>> >>>> Ovidio Marinho Falcao Neto >>>> Web Developer >>>> [email protected] >>>> [email protected] >>>> ITJP - itjp.net.br >>>> 83 8826 9088 - Oi >>>> 83 9334 0266 - Claro >>>> Brasil >>>> >>>> >>>> >>>> 2012/6/27 Massimo Di Pierro <[email protected]>: >>>> > I would suggest you use a different database (a different connection >>>> > string) >>>> > for each hospital depending on the hostname. >>>> > >>>> > <hospital>.domain.com >>>> > import re; regex = re.compile('.*\://(.*?)/.* >>>> > db = DAL('postgresq://..../%s' % >>>> > regex.match(request.env.http_host).group(1)) >>>> > >>>> > In this way you keep your data separate (which is better for security) >>>> > and >>>> > will allow you scale horizontally by setting up multiple db server >>>> > instances. >>>> > >>>> > >>>> > >>>> > On Wednesday, 27 June 2012 18:07:44 UTC-5, Ovidio Marinho wrote: >>>> >> >>>> >> I am making an application to control government medical appointments >>>> >> in more than fifty hospitals all use a single postgres database, but >>>> >> each hospital must have administrator access, Operator and User, and >>>> >> each access must be made with the view of their unique data . >>>> >> >>>> >> is this. >>>> >> >>>> >> >>>> >
