I am running a program that has about 16 -READs in it. I am in the testing process so I'm constantly running the report over and over again to test the modifications that I made. Every so often it bombs out on the first record and I get the same error message each time it bombs out that I pasted below. When I get that message, I just run the program again and it runs ok. I also pasted the filedef I have defined. Thanks.

 

-SET &HLDMRS3   = &FOCCPU || '_' || 'p320b';
-SET &HLDFILE3  = &HLDMRS3 || '.ftm';
-SET &HLDMAS3   = &HLDMRS3 || '.mas';


AFTER MATCH HOLD AS 60_p320b       OLD-OR-NEW
 END
 -RUN
 0 NUMBER OF RECORDS IN TABLE=       56  LINES=      6
 0 NUMBER OF RECORDS IN TABLE=       56  LINES=      6
 0 NUMBER OF RECORDS SELECTED=        6  LINES=      6
 0 NUMBER OF RECORDS SELECTED=        6  LINES=      6
 0 LINES OF MATCH OUTPUT     =       12
 0 ERROR AT OR NEAR LINE    626  IN PROCEDURE ddc_std_summary_ov
 (FOC36219)  AN ERROR OCCURED WHEN OPENING FILE: 60_p320b
 

This message has been edited. Last edited by: Michele Brooks,

May I ask why you are using &FOCCPU ?

This variable should contain the cpu time this session has accumulated.

I assume you only create HLDMRS3 once ?

Also, are you reading the results of a MATCH ?, the datesa and numbers will be stored as binary not alpha.

What are the format of the fields being written to the file ?

like Waz says,
you gotta stick a FORMAT ALPHA in the HOLD statement in order to -READ anything from it.

I am doing an ON TABLE HOLD FORMAT ALPHA. See code pasted below. I am setting up the hold files according the production setup that was established before I started my temporary assignment with this compnay. Hold files use the &FOCCPU variable in their hold file names. My apologies for not getting back to this sooner.

TABLE FILE &HLDMRS3
SUM ID 
BY ID_NAME
WHERE ID_NAME NE ''
ON TABLE HOLD AS &HLDMRS4 FORMAT ALPHA
END
TABLE FILE &HLDMRS4
LIST ID ID_NAME
ON TABLE SAVE AS &HLDMRS5
END
-RUN
-TYPE &LINES


-READ &HLDMRS5 &RECORD_COUNT1.I5. &ID_DATA.A71.
-CLOSE &HLDMRS5
-SET &REC_CNT1 = EDIT(&RECORD_COUNT1,'$$$$9');
-SET &ID1   = EDIT(&ID_DATA,'$$$$$$999999999');
-SET &ID_NAME1 = EDIT(&ID_DATA,'$$$$$$$$$$$$$$$$$$$$$99999999999999999999999999999999999');
-TYPE &RECORD_COUNT1
-TYPE &ID1
-TYPE &ID_NAME1

This message has been edited. Last edited by: Michele Brooks,

quote:

I am doing an ON TABLE HOLD FORMAT ALPHA.

You are doing that when creating &HLDMRS4, but not on the creation of &HLDMRS3 which seems to be the one failing.

quote:

-SET &HLDMRS3   = &FOCCPU || '_' || 'p320b';
AFTER MATCH HOLD AS 60_p320b       OLD-OR-NEW
 END
 -RUN
...
 0 ERROR AT OR NEAR LINE    626  IN PROCEDURE ddc_std_summary_ov
 (FOC36219)  AN ERROR OCCURED WHEN OPENING FILE: 60_p320b

That "AFTER MATCH HOLD" is still creating a binary file.

What would I have to do in order to do a clean -READ without that error message popping up?

1. With &ECHO=ALL, you might be able to narrow down what line of code is throwing the error.

2. In general, I don't use ddnames that start with a numeric.

Hi Michele,

Do you do a -READ directly off you "&HLDMRS3" file? i.e.
-READ &HLDMRS3 ...
If you don't then it is very possible that the issue isn't with your -READs.

You said:

quote:

I also pasted the filedef I have defined.

I don't see it. Please paste it again. It may be the culprit.

And then to echo what @j.gross said - put a
-SET &ECHO=ALL;
right at the top of the procedure, and put -RUN after any JOINs and after each END statement.
That way you can force the execution of each step and possibly have better luck locating the real culprit!

I will try when I get a chance. I'll have to keep running it until I receive the error message. It happens sporatically. Thanks for the feedback. Will keep you posted.

I found out what was causing the -READ error. The BY fields in the first record of the HOLD files were blank. I replaced the blank with dummy data and ran the report several times. I ran the report successfully and did not get a -READ error. I will mark this post [SOLVED]. Thanks to all who provided feedback.