|
View previous topic :: View next topic
|
| Author |
Message |
Dhineshpondy7
New User
Joined: 07 Aug 2019
Posts: 1
Location: India
|
|
|
|
| maximum length of LRECL parameter is 32760. I want to increase the file length to more than that.. Is it possible to increase the length. If so please suggest your thoughts. |
|
| Back to top |
|
 |
sergeyken
Senior Member
Joined: 29 Apr 2008
Posts: 2024
Location: USA
|
|
|
|
| Dhineshpondy7 wrote: |
| maximum length of LRECL parameter is 32760. I want to increase the file length to more than that.. Is it possible to increase the length. If so please suggest your thoughts. |
The dataset size (or "length") depends on the total number of records (or lines, or rows) written to this dataset. It has nothing to do with the limitations of parameter LRECL
RTFM |
|
| Back to top |
|
|
steve-myers
Active Member
Joined: 30 Nov 2013
Posts: 917
Location: The Universe
|
|
|
|
No, it is not possible to increase the LRECL parameter beyond 32760. The binary form of the value must fit into a 2 byte number. I do not know why it is limited to 32760 rather than 32767.
As Sergeyken has noted, there is no restriction on the number of records in a data set, but obviously you seem to think you need a longer record. |
|
| Back to top |
|
|
sergeyken
Senior Member
Joined: 29 Apr 2008
Posts: 2024
Location: USA
|
|
|
|
| steve-myers wrote: |
No, it is not possible to increase the LRECL parameter beyond 32760. The binary form of the value must fit into a 2 byte number. I do not know why it is limited to 32760 rather than 32767.
As Sergeyken has noted, there is no restriction on the number of records in a data set, but obviously you seem to think you need a longer record. |
I can only add that any requirement to have LRECL>32K usually proves to myself at probability 99.99999% that the initial design is somehow wrong.
Normally/smart designed system would never require such stupid long records.
So, instead of looking for a sophisticated workaround to bypass this limitation I dare to suggest: please, review the original design of your product in order to get rid of stupid requirements. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3053
Location: NYC,USA
|
|
|
|
| Quote: |
Is it possible to increase the length. If so please suggest your thoughts.
|
No, for z/OS Data Sets.
Other Options |
|
| Back to top |
|
|
steve-myers
Active Member
Joined: 30 Nov 2013
Posts: 917
Location: The Universe
|
|
|
|
| sergeyken wrote: |
...
I can only add that any requirement to have LRECL>32K usually proves to myself at probability 99.99999% that the initial design is somehow wrong.
Normally/smart designed system would never require such stupid long records. ... |
In general, I agree with sergeyken. However, look at IBM's SMF type 30 record and look at what IBM does to get around the 32K restriction. I suspect Advanced Function Printing in IBM potentially has this problem, though - and I admit I am far from an expert in my knowledge of this data stream - where one page of printed output corresponds to one logical record you potentially have this problem in pages with graphic output.
After some thought I realize the TS is concerned about XML data strings, which can reach improbably long lengths. I believe it was Mr. Sorichetti mentioned, there are methods to deal with these strings, though I also feel these strings run a real danger of getting involved in the buffer overruns that many people regard as dangerous. |
|
| Back to top |
|
|
|
|
