There have always been limits on how big or complex a selection statement can be. And as best I know, these things are not configurable. So... you have to be creative on how you select the items.
For example, if you know the keys to the records of the customers to select - or can build the keys based on the client number - you should be able to take that list and use a FORMLIST to select those records directly. On the other hand, if the CLIENTNO field is a data field in the record and a SELECT is absolutely necessary, you may need to create some preprocessing code that selects all of the records in the file and if a record matches your critieria, write it to a work file or a list in memory that can then be selected or FORMLISTed. With SUBR(..) type dictonaries, the possibilities are actually pretty remarkable. For example, create a routine that takes the list of CLIENTNOs that you know and have those values put into a named common variable. Create a SUBR(..) type dictionary using the same named common that returns a Y/N as to whether that record is a candidate based on the list in common. Then change your select to select any records where that field is "Y". The select command is well within limits, and you can put as many millions of data elements in that list that you want, performance notwithstanding. -K ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
