I have a complex report that in the end gets summarized, and I need to make the report drillable. The data that needs to be displayed when the report is drilled on is already compiled in an earlier hold file. Since it's an FML report, using the SET EXPANDABLE is not an option. Is there someway that I can save the hold file with all of the necessary detail, reference it if the user drills down in the report, and then delete that "hold" file at some later point so it isn't saved indefinitely? Another issue I see is that more than one user may be using this report at the same time with different parameters so when they drill they would have to be able to drill into the "hold" file that was created when they ran the report initially.

Any suggestions, ideas?

Bethany

This looks a bit like a datawarehouse table to me.
IE: create a procedure that runs, lets say, every night, which fills this 'hold' file for you and your users. Any drill down at any time will be directed to this hold file, and it will always be up to date, no need to delete it or something like that.
You could use the windows scheduler (the tasks application of Win2k) to schedule the load of the intermediate file.

Hope this helps.

There are several different ways to accomplish this, the most common being a technique using specific naming conventions for hold files and application directories and, as you mention, cleaning up the directories. There are numerous threads relating to saving hold files on the forum that you can find using the search utility.

Bethany,

darin is right that there are a number of disucussions about this already on the forum. However, I recently posted a technique which held a PDF file that was drilled to from an HTML output. Here's the thread. you can alter it to create a HOLD file instead of a PDF. This isn't a complete solution but it will get you started. Do a search to find other solutions.

Possibly Helpful Technique

It won't help you right now, but 7.6.5 has a FOCCACHE feature which will allow you to save the drill-to data in a temp directory and then automatically clean up after a certain amount of time. The file is only valid during the same browser process. We currently use a technique that creates a temp directory based on the Unix process id or pid. Maybe this will give you some ideas.

Don't forget the POST capabilities with FML (EMR, FRL, what the heck)

T

Tony, didn't it start as SML ?

Bethany,

This is not exactly what you wanted but maybe you can use the idea:

  
-* File bethany3.fex
-DEFAULT &DRILL=NO
-IF &DRILL NE NO GOTO #YES;
TABLE FILE CAR
SUM SALES BY COUNTRY
ON TABLE SET STYLE *
TYPE=DATA, COLUMN=COUNTRY, FOCEXEC=BETHANY3(DRILL=YES COUNTRY=COUNTRY),$
ENDSTYLE
END
-GOTO #OUT
-#YES
TABLE FILE CAR
SUM SALES BY CAR
IF COUNTRY EQ '&COUNTRY'
END
-#OUT

quote:

didn't it start as SML?

Warren,

Yes, It started life as Small Modelling Language, then Financial ML, then Extended Matrix Reporting before becoming Financial Reporting Language. Don't ask for dates as that just causes grey hair!!

I don't think I've left any out but I'm sure Gerry will correct me if I have .

T

Bethany,

How about writing the output file to an Application folder, set aside for the sole purpose of holding temp files?

If you use a combination of date and user id variables to construct the filename, prefixed with a report identifier(?) that should enable you to target the correct file. Then set up an overnight job to remove those temporary files.

I've seen this idea used to great effect in my travels.

T