DSXchange

You need to find out what this character is. Most likely it is either tab char(9) or CrLf char(13) or char(10). You can strip these out with convert(), change() or oconv() functions. Do a search these have been covered a lot.

If you are in the directory where you want to setup an account then just type uv. The $DSHOME/bin should be a part of your PATH variable. This will prompt you do you want to create an account. An account consists of 4 files VOC, D_VOC, VOCLIB and D_VOCLIB. At the TCL prompt type CLEAN.ACCOUNT. This ...

awk or sed would work fine. I will ask Peter my AIX expert. I never knew this did not work on AIX. AIX is based on a really old version of UNIX. Did you try it as ksh?

If you go to my web site below or ADN you can download jobs which will get all row counts from all your jobs unless you just want to learn all this. EtlStats.zip has the dsx and all the create table scripts to post these to an Oracle table but you can easily switch this to any database.

That is what I was trying to explain. Poor explaination on my part. I will try dsenv.

I noticed that when we ran jobs from the scheduler that the umask was 022. We checked and umask 002 was set in the ds.rc file. If you ran the job with the at command then it was set to 002. Only when you ran the job in crontab was it 022. My fix was to create a UV.LOGIN in the VOC of the UV account....

Do it in a shell script. I never use the ftp stage. Too hard to debug. Stays connected longer.

When you route data from one transform to another transform then that will slow a job down more than 15 lookups on one transform. If you cannot combine these into one transform then let us help you.

Build it both ways and let us know which is faster.

You can use dsjob to execute any job from a UNIX or DOS prompt. There are lots of posts for doing this. Most use batch files or shell scripts. Do a search. Ken Bland has a nice shell script he posted long time ago.

Ray Almost all UNIX systems the set command by itself works like env. It will display all the shell variables and their values. The set command the way I use it will parse a shell variable ignoring multiple spaces and load $1, $2 and so on. I think this will work on AIX. I am used to use it. I prefe...

DataStage does lots of data type conversion automatically. Leave the metadata on the source and the targets alone. Just import them from the database of your choice. If you route data from one column to the target normally it will do whatever conversion is necessary behind the scenes to get the data...

The dsjob.exe command I think is on the CD that you install the DataStage client from. I am not sure if you can get it without installing a DataStage Windows server. If you do not have a dsjob then all it does is create a list of all the jobs. You can do this on your UNIX server. The dsjob command c...

Ray I have used this technique for years. I do not remember if it works on AIX. If not then use a cut command. The way I used this set command will parse a shell variable value into $1, $2, $3 and so on. If you know a better way then explain it. This is just a quick answer to give them an idea. Ther...

You can do this in the shell script. What line is the date on? With tail and head you can extract that line. You can parse the line with the set command to get the second or third word if that is the date.

It should create the folder using KgdMakeDir routine. Sometimes it does not work. Manually create a KimD directory below your project. Below that create a Jobs directory and a Routines directory. If KgdMakeDir is not working then you may need to create 20040907 directory. ProjectDir/KimD/Jobs/200409...