Variable Length Input File
in [Cobol]
|
From: Anonymous on 26 Dec 2006 18:11 In article <2cafe$459199e9$d066072d$978(a)FUSE.NET>, CG <Carl.Gehr.ButNoSPAMStuff(a)MCGCG.Com> wrote: >docdwarf(a)panix.com wrote: [snip] >> Dealing with a VB dataset with different-length records has not presented, >> at least to me, very much of a problem... the difficulty is, I believe, in >> dealing with a fixed length dataset which has different-length records. [snip] >???? > fixed length dataset which has different-length records. ???? > >That is a total contradiction! Either the file has FIXED LENGTH records >or the records VARY in length. You might want to inform the person who started this thread about that... but the last heard was from that account in this thread seems to have been as follows. From <http://groups.google.com/group/comp.lang.cobol/msg/e2bd0f47c45850c8?dmode=source> --begin quoted text: I think I didn't explain my problem well. Maybe it'll be better with an example. Let's say, I have file called TESTA. TESTA doesn't always have the same LRECL. Sometime TESTA will be defined in the JCL with LRECL=30. Other times, it will be have LRECL=70. And so on. My problem is how should TESTA be represented in FD? --end quoted text I was able to find the entire thread beginning at <http://groups.google.com/group/comp.lang.cobol/browse_thread/thread/5f5078dee2ab20f1/1418ce24a2de8ee3?q=%22BLOCK+0+and+the+actual+record+length%22&lnk=ol&> DD
From: Pete Dashwood on 26 Dec 2006 18:54 I have a question, please, Doc. (It is many years since I worked in this environment (actually, it was an environment LIKE this one because this environment was not then available, if you catch my drift...)) I have retained enough of what I learned there to understand the comprehensive samples you posted, and note in passing that nothing much appears to have changed as far as JCL is concerned... I too, would have expected RECFM=VB, but I understand why you didn't put that. My question is this (and forgive my ignorance if there is some obvious reason that I simply haven't tumbled to): If the JCL specifies a block size of 27990 (nicely divisible by 30, 933 times), and the COBOL program is specifying a block/record size of 32767, wouldn't this be likely to cause an IO error when the dataset is opened? Just a thought... Pete. TOP POST - nothing more below <docdwarf(a)panix.com> wrote in message news:emsa6g$n0j$1(a)reader2.panix.com... > In article <2cafe$459199e9$d066072d$978(a)FUSE.NET>, > CG <Carl.Gehr.ButNoSPAMStuff(a)MCGCG.Com> wrote: >>docdwarf(a)panix.com wrote: > > [snip] > >>> Dealing with a VB dataset with different-length records has not >>> presented, >>> at least to me, very much of a problem... the difficulty is, I believe, >>> in >>> dealing with a fixed length dataset which has different-length records. > > [snip] > >>???? > fixed length dataset which has different-length records. ???? >> >>That is a total contradiction! Either the file has FIXED LENGTH records >>or the records VARY in length. > > You might want to inform the person who started this thread about that... > but the last heard was from that account in this thread seems to have been > as follows. From > <http://groups.google.com/group/comp.lang.cobol/msg/e2bd0f47c45850c8?dmode=source> > > --begin quoted text: > > I think I didn't explain my problem well. Maybe it'll be better with > an example. Let's say, I have file called TESTA. TESTA doesn't always > have the same LRECL. Sometime TESTA will be defined in the JCL with > LRECL=30. Other times, it will be have LRECL=70. And so on. My > problem is how should TESTA be represented in FD? > > --end quoted text > > I was able to find the entire thread beginning at > <http://groups.google.com/group/comp.lang.cobol/browse_thread/thread/5f5078dee2ab20f1/1418ce24a2de8ee3?q=%22BLOCK+0+and+the+actual+record+length%22&lnk=ol&> > > DD >
From: Anonymous on 26 Dec 2006 19:56 In article <4vdr12F1botahU1(a)mid.individual.net>, Pete Dashwood <dashwood(a)removethis.enternet.co.nz> wrote: >I have a question, please, Doc. [snip] >If the JCL specifies a block size of 27990 (nicely divisible by 30, 933 >times), and the COBOL program is specifying a block/record size of 32767, >wouldn't this be likely to cause an IO error when the dataset is opened? I thought there would be some kind of error as well, Mr Dashwood, hence my testing... but both of our thinkings appear to be contrary to Mr Morris' assertion that 'For FB code the maximum size 01 level expected for the FD and RECORD 0 in addition to BLOCK 0 and the actual record length will be determined at run time. Unfortunately this only works for INPUT.' I posted code which I believed was in accordance with this - it is a new technique for me and I'm not quite sure I got it right - and, likewise, posted the results (an ABEND). If there's something else I should be doing then, by all means, post the code - both COBOL and JCL - and I'll be more than happy to compile it and show the results. DD
From: CG on 26 Dec 2006 19:57 docdwarf(a)panix.com wrote: > In article <2cafe$459199e9$d066072d$978(a)FUSE.NET>, > CG <Carl.Gehr.ButNoSPAMStuff(a)MCGCG.Com> wrote: >> docdwarf(a)panix.com wrote: > > [snip] > >>> Dealing with a VB dataset with different-length records has not presented, >>> at least to me, very much of a problem... the difficulty is, I believe, in >>> dealing with a fixed length dataset which has different-length records. > > [snip] > >> ???? > fixed length dataset which has different-length records. ???? >> >> That is a total contradiction! Either the file has FIXED LENGTH records >> or the records VARY in length. > > You might want to inform the person who started this thread about that... > but the last heard was from that account in this thread seems to have been > as follows. From > <http://groups.google.com/group/comp.lang.cobol/msg/e2bd0f47c45850c8?dmode=source> > > --begin quoted text: > > I think I didn't explain my problem well. Maybe it'll be better with > an example. Let's say, I have file called TESTA. TESTA doesn't always > have the same LRECL. Sometime TESTA will be defined in the JCL with > LRECL=30. Other times, it will be have LRECL=70. And so on. My > problem is how should TESTA be represented in FD? > > --end quoted text > > I was able to find the entire thread beginning at > <http://groups.google.com/group/comp.lang.cobol/browse_thread/thread/5f5078dee2ab20f1/1418ce24a2de8ee3?q=%22BLOCK+0+and+the+actual+record+length%22&lnk=ol&> Sorry, I jumped in based on your examples and the "Subject" line... The 'input file' then, contains FIXED length records, but the program needs to anticipate the length of those differing fixed length files. [Hmmm, easy to handle in PL/I.] Unless the records are somehow self-defining so ODO can be used in a single record definition, I think COBOL is going to always assume V format if it sees multiple, different length record definitions. [Ref back to Bill Klein's post on 12/21/06 05:36 pm.] I'll butt out... CG
From: Anonymous on 26 Dec 2006 20:11
In article <18530$4591c500$d066072d$7705(a)FUSE.NET>, CG <Carl.Gehr.ButNoSPAMStuff(a)MCGCG.Com> wrote: >docdwarf(a)panix.com wrote: >> In article <2cafe$459199e9$d066072d$978(a)FUSE.NET>, >> CG <Carl.Gehr.ButNoSPAMStuff(a)MCGCG.Com> wrote: >>> docdwarf(a)panix.com wrote: >> >> [snip] >> >>>> Dealing with a VB dataset with different-length records has not presented, >>>> at least to me, very much of a problem... the difficulty is, I believe, in >>>> dealing with a fixed length dataset which has different-length records. >> >> [snip] >> >>> ???? > fixed length dataset which has different-length records. ???? >>> >>> That is a total contradiction! Either the file has FIXED LENGTH records >>> or the records VARY in length. >> >> You might want to inform the person who started this thread about that... >> but the last heard was from that account in this thread seems to have been >> as follows. From >> ><http://groups.google.com/group/comp.lang.cobol/msg/e2bd0f47c45850c8?dmode=source> >> >> --begin quoted text: >> >> I think I didn't explain my problem well. Maybe it'll be better with >> an example. Let's say, I have file called TESTA. TESTA doesn't always >> have the same LRECL. Sometime TESTA will be defined in the JCL with >> LRECL=30. Other times, it will be have LRECL=70. And so on. My >> problem is how should TESTA be represented in FD? >> >> --end quoted text >> >> I was able to find the entire thread beginning at >> ><http://groups.google.com/group/comp.lang.cobol/browse_thread/thread/5f5078dee2ab20f1/1418ce24a2de8ee3?q=%22BLOCK+0+and+the+actual+record+length%22&lnk=ol&> > >Sorry, I jumped in based on your examples and the "Subject" line... >The 'input file' then, contains FIXED length records, but the program >needs to anticipate the length of those differing fixed length files. This is what I believe to be the case, aye... once again, the original poster's comments are few and far between. DD |