Re: Multiple instances of SE?

From all I know and have tested there is no support for multiple instances of the SE engine with access from one to the other via insert into orders -- in local garpacse-/dbs/yutzdb select * from yutzdb@javase.orders -- The one in /jvdbs/yutzdb or any similar syntax. As you say there are some fundamental problems as to what multiple instances would mean and where the databases would be located. You can change from one database to another within a program though and there are some tricks that have worked using temporary tables but I have never tested them and don't know when or even if they work. However as far as I remember the idea was to select into a temporary table first. Then switch database and insert from that temporary table. It should be easy enough to test and might be simpler than unload files. On Fri, 31 Jul 1998 16:43:29 GMT, Jacob Salomon <jake@garpac.com> wrote: >Hi Family, > >is it possible to have multiple instances of SE defined on one SQLHOSTS >file? I blindly started set up for a second instance as follows: > >garpacse seipcpip scosysv sqlexec >javase seipcpip scosysv duuuuhh? > >Say, what *would* I put in the service name for a second SE instance? Or >does it not really need to be sqlexec? > >My situation is that I want the database name (yutzdb for the purposes >of examples) to be the same on both "instances". > >I know I can change which yutzdb database a program is referencing by >playing with the DBPATH environment variable. I might have a yutzdb >database under directory /dbs and another yutzdb database under >directory /jvdbs; by setting DBPATH to one of these directory names >(before starting the program) I can be certain of which yutzdb database >I am referencing. This will solve the basic existence problem. > >Now I want to be able to copy data from one of these databases to the >other, prefereably without the intermediate .unl files. I need to be >able to do it in a single SQL statement - > insert into yutzdb.orders -- in local /dbs/yutzdb database > select * from yutzdb.orders; -- from /jvdbs/yutzdb database> >(Yes, I suppose I would have to run two instances of sqlexecd in order >for this to work.) > >Hence, my desire to use the more familiar syntax with server names as >part of the table names: > > insert into orders -- in local garpacse-/dbs/yutzdb > select * from yutzdb@javase.orders -- The one in /jvdbs/yutzdb> > >Is such a cross-server addressing scheme even possible in SE? How would >the server name (garpacse, javase) even relate to the different >directories containing the respective databases? > >Or am I cluelessly off the deep end on this one? (I realize some of my >questions are based on the dubious assumption that it is possible.) > >Thanks much. >-- > -- Jake (Retrospectively realizes there is no future in hindsight) >+------------------------------------------------------------+ >| The expedient performance of a task with excessive concern | >| regarding its duration-to-completion engenders a virtual | >| certainty of diminished benefit therefrom. | >| -- Benjamin Franklin (but he said it in 3 words) | >+------------------------------------------------------------+ Nils Myklebust NM Data AS Norway E-mail: Nils.Myklebust@nmdata.com FAQ at: http://www.iiug.org/techinfo/faq/faq_top.html (Now with ODBC info under "Third party products".)