Re: About pyvertica (a clone of pyscopg2)
От | Daniele Varrazzo |
---|---|
Тема | Re: About pyvertica (a clone of pyscopg2) |
Дата | |
Msg-id | CA+mi_8baXqinV=V_oUw6VENqzCid50TnvOT9-NZFwZ3jPNtGOw@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: About pyvertica (a clone of pyscopg2) (Federico Di Gregorio <fog@initd.org>) |
Ответы |
Re: About pyvertica (a clone of pyscopg2)
|
Список | psycopg |
On Wed, Sep 7, 2011 at 10:12 AM, Federico Di Gregorio <fog@initd.org> wrote: > Hi Angel, > > I hope you don't bother I added the psycopg mailing list in cc: other > people can be interested in your work. :) > > On 06/09/11 20:22, Angel Freire wrote: >> Hi, my name is Angel and since a few weeks now I've been working with >> some database engine called "Vertica" it's a column-oriented db based >> on PostgreSQL, it's so based on PostgresSQL that with a few little >> modifications pyscopg2 is able to connect to it and perform queries >> without any problems. I've found the patch needed to make pyscopg2 run >> with Vertica in StackOverflow >> (http://stackoverflow.com/questions/2469167/merge-excel-cells-using-pyexcelerator). > > I can't find a patch there. :( It may have been this: http://stackoverflow.com/questions/2695684/using-an-odbc-application-with-a-jdbc-driver Looks like Vertica has a few differences in the types oid. Unfortunately there is at least a conflict: from the patch I see oid 16 is used by Vertica as a decimal type, while Postgres uses the same oid for the bool type. A solution could be to detect vertica at connection time (which seems done by checking server_version = 0, I don't know if there is a more robust way) and re-organize the typecaster map accordingly. -- Daniele
В списке psycopg по дате отправления: