Hi,

I have a user trying to merge two hold-files together via InfoAssist. Considering MORE is not supported by InfoAssist, I'm trying to come up with an alternative approach to get the same result, but failing at it.
Anybody out there with an idea?

Hold-file1:

 
| YEAR | TYPE  | AMOUNT|
| 2013 | TYPE1  | 1 |
| 2013 | TYPE2  | 2 |
| 2013 | TYPE3  | 3 |
 

Hold-file2:

 
| YEAR | TYPE  | AMOUNT|
| 2013 | TYPE4  | 4 |
 

Intended report:

 
| YEAR | TYPE  | AMOUNT|
| 2013 | TYPE1  | 1 |
| 2013 | TYPE2  | 2 |
| 2013 | TYPE3  | 3 |
| 2013 | TYPE4  | 4 |
 

This message has been edited. Last edited by: <Kathryn Henning>,

Hi Peter,
What type of data file are you reading?

Peter,
Have you tried to join the two hold files? You might need to save as format Focus and index the field used in the join.

Luiz

@Twanette: source = database, stored in intermediate FOCUS hold file. (goal is not to intervene on database level)

@Luiz: I don't see how a join could result in mentioned result?

Hi Peter,

I agree with you that you probably wouldn't be able to "JOIN" the data files - you need to "concatenate" them.

I don't think one can do it directly in InfoAssist (well, I haven't managed to find a way).
So, the idea that I had was that perhaps one could issue either a FILEDEF or a USE command that would tell WebFOCUS to concatenate the HOLD files.

But since once cannot do this within InfoAssist, one would need to issue this command before the time e.g. in a PROFILE e.g. in EDASPROF.PRF

The problem with this approach though is that it enforces certain rules/standards that the user would need to follow to get it to work.

But let me explain the ideas - perhaps it gives you a few things to try.

To dynamically concatenate "text" files with a FILEDEF command, you would issue a FILEDEF with a wildcard e.g.
FILEDEF SUMMDATA_1 DISK SUMMDATA_*.FTM
The above would use a master file called SUMMDATA_1, and would read all FTM files in the user's temp directory that have a filename that starts with SUMMDATA_
So you would then HOLD the data files as text files with the names SUMMDATA_1, SUMMDATA_2 etc.
This method does not physically concatenate the files - it happens dynamically.

Or, you could issue a FILEDEF with an (APPEND option. e.g.
FILEDEF SUMMDATA_1 DISK SUMMDATA_1.FTM (APPEND
Then you would HOLD each file as a text file, but with the same name i.e. SUMMDATA_1
Obviously this will only work if InfoAssist does not re-issue or clear FILEDEFs before creating the HOLD files.
The data should then physically be written to the same FTM files.

Or, for FOCUS HOLD files, perhaps you could issue a USE command. The problem with this is that you will need to know how many interim HOLD files will be created e.g.
USE
SUMMDATA_1.foc AS SUMMDATA_1
SUMMDATA_2.foc AS SUMMDATA_1
END
What I don't know is what InfoAssist will do if there is a USE command for a FOCUS file that still needs to be created, and whether you can use the "NEW" option in a concatenation.

After all that, the final report would use SUMMDATA_1.

Just a caution though - I haven't tried any of the above ideas - they're just ideas ;-)

Twanette,

If I understand you correctly, the edasprof FILEDEF entry would pretty much merge all SUMMDATA_* hold files in a user's temporary directory whenever a report is executed.

Two things I don't understand are:
- Is the merged data then available in SUMMDATA_1? (i.e. I can use SUMMDATA_1 as source for my report in IA and it will contain all data during execution)
- Wouldn't the edasprof merge be executed before the actual hold files are created? (as I understood, first edasprof is executed, and then the report code)

Thanks for the input - much appreciated!

Hi Peter,

As I said "I haven't tried any of the above ideas - they're just ideas"

And the two things you "don't understand" are actually correct as you stated them.

So I am really hoping that we can somehow "fool" InfoAssist.

If it works as I hope, then:
- the FILEDEF that gets issued in EDASPROF will still be active after the HOLD files in the temp directory were created
- the FILEDEF in EDASPROF won't interfere with the first HOLD file (SUMMDATA_1) that gets created, because we need to use the SUMMDATA_1 master file to read the data.

Obviously all of this would be so much better if one could include a "-INCLUDE" in an InfoAssist process. (I don't think you can).

Then you would first create your HOLD files (SUMMDATA_1, SUMMDATA_2 etc.), and then -INCLUDE a stub program that issues the FILEDEF, and then paint/run your report.

Aha - got it!

I don't know when I'll be able to test this out (time being a precious commodity these days), but I'll make sure to update the post with my results.

Thanks for the help!

MFD_PROFILE could be leveraged to perform the data setup (issuing the USE or FILEDEF).