vasukb

 Profile Answers by vasukb 

• Feb 14th, 2006

 

well, it depends on whether the data will be accessed directly or sequentially.

for direct access use small ci

for sequential access use larger ci.

Culver_lake

 Profile Answers by Culver_lake 

• Mar 18th, 2006

 

Don't get bogged down unless the file is tremendously large. Rule of thumb: use 4096 for the data and 1024 for the index. You'll never never be far from optimum with those CI sizes.

kay kay

• Sep 13th, 2006

 

Please let me know why we choose data as 4096 and index 1024 in

are they multiples,if not then why not 4000 or 1000

explain in details,Please?

ragavprasanna

 Profile Answers by ragavprasanna 

• Sep 1st, 2008

 

Control interval size can be optimally selected by the system when we omit it. Because the C.I. is nothing  but the place where a record get stored, depending on various factors (such as either a fixed length or variable length record) we have to calculate the C.I size. Don't worry about these issues, thanks to some modern developments by the IBM which makes the system to effectively calculate the optimum C.I. size.

dhs01

 Profile Answers by dhs01 

• Apr 15th, 2009

 

1024 and 4096 are powers of 2. You could specify 1000 and 4000 but VSAM would actually use 1024 and 4096 although I'm not sure if the full 1024 would actually be available to you. Short answer is that since that is the number VSAM will use, 1000 or 4000 will not save you anything.  

nlarsensmith

 Profile Answers by nlarsensmith 

• Dec 1st, 2010

 

Best way is to choose small CI size for random access and large CI size for direct access.

Without specifying CISZ the defaults will be 4096 for data and 1024 for index.

Remember these sizes are obtained because 2 raised to n gives them. The standard is to specify 512b, 1024b or 1k, 2k and 4k.Size within 1k should be multiples for 512(512*2), if it is to be more than 1k it should be a multiple if 2(1k*2=2k and 2*2=4k), but not more than 4k.