Joined: 23 Nov 2006 Posts: 19270 Location: Inside the Matrix
According to SUPERC, this is the cause of a 26:
26 - Error. The listing data set could not be extended and
condition). The results are truncated and incomplete.
However, this is what they say about a 28:
28 - Error. No data was compared because the old and new data sets were empty, because member names were invalid, or because there was nothing left to compare after input filtering. See the listing for more
SUPERC compare step gives return code [value yet to be confirmed] when i/p fiels [sic] are empty. How can this be avoided. Is there any parameter to treat empty i/p fiels as equal and get return code zero.?
Have you tried manual, online help, google?
I assume you are in batch. Why are the "files" empty? If you have a dataset which can be genuinely empty, how do you know that any given case of an empty dataset is not an errorneous one? Errorneous and genuine are identical (empty).
I don't like empty datasets existing on purpose. I like headers and trailers. Business date and record count/hash count. I know the file is genuine, is from the right day, and contains zero data records. Anything I'm comparing it to which should be the same, will be the same.
I also don't have to worry about problems like yours, or what to do if I get EOF on a priming read, etc.
I sympathise with your problem (I'm assuming it was not down to you originally), but a better system design would have avoided this and similar problems.