Why would I get the following message? I thought that was the strength of WebFOCUS - being able to join tables from multiple databases/nodes/subsystems?

(FOC2510) FOCUS-MANAGED JOIN SELECTED FOR FOLLOWING REASON(S):
(FOC2519) THE JOINED SEGMENTS RESIDE IN MORE THAN ONE NODE OR SUBSYSTEM
(FOC2590) AGGREGATION NOT DONE FOR THE FOLLOWING REASON:
(FOC2592) RDBMS-MANAGED JOIN HAS BEEN DISABLED
SELECT T1.Con_GUID,T1.Fund_Number,T1.DCRC,T1.Fund_Name,
T1.City,T1.FLName,T1.MYNS,T1.YYNS,T1.AYAS FROM
EIDWCRM.dbo.CRM_TotalsView T1 WHERE (T1.Con_GUID =
'{5BAAFAFF-6D2E-4634-9E8F-B6C8E6B067E1}') AND
NOT((T1.Fund_Number IN('421', '422', '423', '424', '425',
'428', '431', '432', '433', '434', '435', '450', '451', '452',
'453', '454', '455', '456', '457', '458', '459', '463',
'490')));
SELECT T2.CF_Percent,T2.CF_CompanyIdName,
T2.StateOrProvince,T2.Country,T2.Telephone1 FROM
GGOF_CRMEXT.dbo.ContactView T2 WHERE (T2.ContactId = ?);

The JOIN is being honored, but...

The first two messages inform you that Focus could not pawn off the job of assembling the joint records to the server, because the data do not all reside on a single server. Instead the logical join is being accomplished by Focus ("FOCUS-MANAGED JOIN SELECTED"), which reads separately from the two sources and assembles the joint records.

Additionally (as the next 2 messages note), as a result Focus had to pull detail-level data from the two servers, rather than letting the server boil down the data to the requested level of aggregation.

In other words, it's doubly inefficient relative to the case where all the data are available through a single server.

This message has been edited. Last edited by: j.gross,

As j.g. says the join is being honored.

When working within a single database structure or databases on the same platform that can talk to each other
then the principle database can manage the process, however even they may need to scan and/or spool.

The issue is when you are cross platform and/or cross database, they often are unable to manage the work and so Focus/WebFocus
takes control and spools the data to the lowest level requested from each database/system and then resolves your request.

Often it is more efficient to request to the level required on the difefernet database/system and then perform your join on the resultant hold files (either dbms or indexed Focus) of cause if you do not have all the required keys for the extracts then it may not be such an advantage.

A method that can dramatically improve performance if the first source is supplying most of the keys for selection from the second is to extract your keys from the first and then pass them as selections to the second and then processing the resultant answer sets.
Or if you are wanting to work from rolled up data at the same level from the different sytems extract the rolled up data and then process that. That way you pass as much of the work as possible to the DBMS.

I've seen these sort of techniques reduce processing times by up to 80 %.

Cross platform/database joins are inherently inefficient, that said Focus/WebFocus is just about your only option unless you want to go to an ETL type based solution which just means that the data is not live but staged and always out of date.

Thanks to j.g. and JG.

(what are the odds?)