View previous topic :: View next topic
|
Author |
Message |
socker_dad
Active User
Joined: 05 Dec 2006 Posts: 177 Location: Seattle, WA
|
|
|
|
I'm working with a SyncSort job, and came across the following line:
Code: |
SORT FIELDS=(1,2,FI,A) |
What is FI? The manual relates it to binary (BI), but doesn't go into any more detail.
The data in the field being sorted is (Zone/Digit format):
I would have used binary (BI), but maybe the FI is better? |
|
Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011 Posts: 7309 Location: Inside the Matrix
|
|
|
|
FI is signed. Left-most bit 1, indicates negative. BI is same thing, but unsigned, left-most bit is part of any value.
If you have signed-binary, and you want the negative values first in the sequence, you'd need FI. If you have unsigned binary, then you should use BI, else a mess can ensue.
So, can your field be negative? Is is from Cobol without TRUNC(BIN), so the maximum value is 9999? Then you'd not notice the difference from the output.
FI may be slower to process as it would have to differentiate between + and -, which BI doesn't. Not of much significance, I'd guess.
I'd also go for BI, unless it is a signed field. Unless from Cobol with TRUNC(BIN), then you'd have more checking-up to do :-) |
|
Back to top |
|
|
socker_dad
Active User
Joined: 05 Dec 2006 Posts: 177 Location: Seattle, WA
|
|
|
|
The field cannot be negative (it's a state/region id number), so it looks like FI and BI are both acceptable choices; however, BI would make more logical sense.
Thanks for the insight! |
|
Back to top |
|
|
|