I am trying to join two tables and the problem is that in one the field is Alpha and in the other its a number. I thought the following would work but it just runs forever and does nothing. What is wrong with this code?

JOIN CLEAR *
-*
DEFINE FILE F4211
NEWFIELD/I8=EDIT(SDRORN);
END

TABLE FILE F4211
PRINT NEWFIELD SDRORN 
WHERE SDDOCO EQ 793581
AND SDRORN EQ '00413249'  
ON TABLE HOLD AS NSR01
END
-RUN



JOIN NEWFIELD IN NSR01 TO WMDOCO IN F3111 AS J1

TABLE FILE NSR01
PRINT NEWFIELD SDRORN WMDOCO 
END

This message has been edited. Last edited by: Kerry,

Is "WMDOCO" (IN F3111) an I8?

What segtype is "WMDOCO" (IN F3111)?

Are you getting anything out of the final TABLE FILE request? Do an ON TABLE HOLD and see the number of LINES and RECORDS... Let us know...

Read up on the JOIN Syntax for more info...

It turns out it does work after all, just takes forever. I find it strange since its only one record, but I guess its normal.

No Josh, That should NOT Be Normal. Perhaps there a cartesian in there someplace...

How many records are in "NSR01"? "F3111"?

Did you do the HOLD thing in my previous post?

Josh,

This is one of those times that it might be advisable to use SQL passthru against your JDE files so that you can achieve the extract in one pass instead of two bites.

T

You may be able to forward-define that column:

JOIN CLEAR *
JOIN NEWFIELD WITH SDRORN IN F4211 TO WMDOCO IN F3111 AS J1
DEFINE FILE F4211
  NEWFIELD/I8 WITH SDRORN = EDIT(SDRORN);
END
TABLE FILE F4211
  PRINT NEWFIELD SDRORN WMDOCO 
END

What's taking so long is probably WebFOCUS downloading both tables to do the join locally. With some luck the above scenario is enough to help WF generate an SQL query for the join instead.