DSXchange

I was on an EE assignment where the number of stages exceeded 1,000 in several jobs. My firm was called in midway through the project, after the jobs were written, to help resolve 'some issues'. After some modifications, and much tweaking of kernel, memory, and disk resources, the jobs did actually ...

Generally, the tr command will yield the best performance when translating character values on a UNIX system.

Carter

Another caveat, my documentation (7.5) specified that a C++ compiler was required.
I have not tried (yet) on a Windows platform.

Carter

OK, 200 input characters (assume a-z,A-Z,0-9, about 64 possibilites?). Each segment is 40 of these characters, so the number of different possible character patterns is: 64^40 = 1.76 x 10^72 Each of these segments is translated by the CRC function into an 8 character string of hex. The number of dif...

Are you writing a routine just so you can append?
Otherwise, use a Sequential file stage, and click 'Append to existing file' on the Inputs tab.

Carter

I needed my microseconds now, so I added an external routine for the purpose: /* getSysUsec.c gets sys seconds and usecs, formats as "%d.%06d", returns string */ #include <unistd.h> #include <stdio.h> #include <time.h> #include <sys/time.h> char * getSysUsec(){ struct timeval time; char re...

Hmm..., that's exactly the point. You said that you were using the CFF stage, right? Have you read the functional description of what a CFF does? In COBOL, the REDEFINES clause allows different layout structures for records contained in the same file. See if you can obtain the COBOL FD for the file ...

You are already using CFF, so what did you use for file metadata? EBCDIC files with headers/trailers typically are defined in terms of a COBOL file structure, where the header and trailer are REDEFINES of the base record structure. I believe that the CFF will handle such correctly, as long as the sc...

You can use the ls -1t method, but some folks have the unfortunate habit when viewing files, of writing a new version (shift ZZ, you know who you are), which resets the UNIX file timestamp. You will be lucky to find out about one of these; more often the wrong file is silently processed, the details...

Arnd,
Did you ever get the microseconds to show?

I'd much rather use a DataStage built-in to get usec, than roll custom code with gettimeofday() etc.

Carter

It seems from your example that the second column of Trailer
records always contains 'TRA'?
So, since you are using a Transformer, why not just add a constraint like this:

col2 <> "TRA"

Carter

Be careful with your assumption that all dates will be in 21 century. The code below will fail with an input date of 99032 (1999-02-01). The lower 3 digits are the day of the year, the remaining upper digits are years since 1900. Carter Hello folks, After 3 days of work with all the iconv,oconv and ...

I believe you must add 1900 to the first result, not 1899:

ex) 105030 = 2005-01-30
Int(105030/1000) = 105
105 + 1900 = 2005

Carter

For dates in years between 1900 and 9999, you can convert in DB2 like this: DB2 internal date = DATE(RTRIM(CHAR( D1 + 1900000 ))) where D1 is the julian-like-date Using your example of 105030, DATE(RTRIM(CHAR( 105030 + 1900000 ))) DATE(RTRIM(CHAR( 2005030 ))) DATE(RTRIM('2005030 ')) -- DB2 creates 1...

I have seen some ftp that suppport umask for the target files. But most ftp implementations support some kind of local shell capability, so you can say something like this in your ftp script after the files have been recieved. .... get $filename .... ! chmod 111 $filename .... Check the documentatio...