DSXchange

Hi,

My source files are complex flat files from V series Mainframe systems which send files in EBCDIC fixed length format.

I collect rejects (using the save reject mode property) from these source files into another sequential file (just being curious). Now what type of rejects should I expect to get collected in the sequential file?

For Example:

Say the source file (converted to ASCII) is like this:

Metadata - 9(3)X(3)X(3)

[ 9 stands for numeric in Mainframe world ; X stands for CHAR]

File Data: [Ignore newlines - Actual files in fixed lenght format]

001AAABBBCCC
002CCCDDDEEE
003CCCRRRTTT
...
XXXYYYZZZQQQ
<blank><blank><blank>EEERRRTTT
.....

Now I belive that I should see 2 records [ the last two shown above] in the reject file. Am I correct?


Also can sombody confirm that if the number of data bits in the file [excluding headers and trailers], are not a multiple of record length, then DataStage will not be able to read the source file [aborting with short read message - Schema read failure]?

Thanks!!

I would expect to see all these rows in the reject file. You have declared three fields (999., XXX. and XXX.) but the "good" data have more fields than three.

ray.wurlod wrote:I would expect to see all these rows in the reject file. You have declared three fields (999., XXX. and XXX.) but the "good" data have more fields than three. ...

Oh.. Very sorry... The original structure was 999,XXX,XXX and XXX.

vrishabhsagar wrote:

ray.wurlod wrote:I would expect to see all these rows in the reject file. You have declared three fields (999., XXX. and XXX.) but the "good" data have more fields than three. ...

Oh.. Very sorry... The original structure was 999,XXX,XXX and XXX.

Anyone has any ideas? Kindly help me with this....

How about you tell us what you get on the reject link from the CFF stage?

ray.wurlod wrote:How about you tell us what you get on the reject link from the CFF stage? ...

Okay let me paraphrase so that my dilemma becomes clearer.

I have a source CFF (with EBCDIC Fixed length data) and I rout the reject link into a column restructure sage and convert the varbinay input into a varchar output (and peek it)

Now I expect to see the whatever is being rejected from the source file in the director logs, but although the number of rejects is 2 (as expected); I can only see '??' in the director logs.

Also when I try to dump the restructured rejects into a seq file, I see blank lines (2 blank lines) in the file.

Is this happening because restructure stage (functionally) does not convert the EBCDIC source data into ASCII before giving a VARCHAR output?