DSXchange

Please DO NOT double post. You have an answer in your other thread with this problem.

Moderator - please delete this duplicate thread.

You convert the dates to the internal format using ICONV(In.DateColumn,'D4/DMY[2,2,4]') . Then subtract the two results to get a number of days difference. If you want to output this number of days starting at 01/01/0000 then you can do OCONV(DateDiff+ICONV('1/1/0000','D4/DMY[2,2,4]'),'D4/DMY[2,2,4]...

The ODBC connectivity to text files exists in UNIX as well, it is just a lot easier to set up a quick system DSN on Windows than on UNIX. I am not at a DS client where I can test this at the moment, but the cartesian product should be easy to implement, although it is going to be quite slow and usin...

It looks like you have neglected to specify a Schema and the mload defaults to a different schema and thereofore the table is not being found.

As you are on Windows, declare the two text files to be ODBC data sources and then you can do your cartesian join on them.

Colin, the purge settings are in the log file in the record //PURGE.SETTINGS. You would need to read this record, decide if the log's settings are for {n} days or {n} runs and then go through all of the RT_LOG files records and remove only those that are applicable. I assume that this is a bit more ...

Kris, each client instance of the 3 main tools (designer,director and manager) can be open to a different project (on the same or another host) at the same time.

The one I mentioned earlier, use Reflink.NOTFOUND builtin value.

Craig - that's odd, when I "edit" my post the code shows up nicely indented, but not in the post. I wonder if I need to use quote tags?
--> Correction, I neglected to close the code tag, so the indentation disappeared

Hmmm... if referential constraints in your table are stopping you from loading data you really only have two options:
1. turn off referential integrity or disable constraints
2. fix your data so that it doesn't violate the constraints.

The problem is not so much hashed file performance as it is job design. Can you not flag changed records in your table so that instead of reloading 80 million rows each and every time you run this job you will just update those values that have changed?

vsi wrote:... Because once i used the comman LIST.READU and all the jobs just vanished...

The LIST.READU command cannot do this, but the DS.TOOLS reindex can do this if you run it while others are working.

I have had no problems using FTP with fixed width files. What seems to be the error?

First off, you cannot TRIM a CHAR field. It always has the same length. You can convert implicitly from CHAR to VARCHAR but it will not trim the spaces. If you are writing to Oracle VARCHAR2 columns the trailing spaces are automagically stripped. But you should do your own explicit trim in a modify ...

Hadi - wouldn't it be easier to design a trivial DS job to do this? Also, PLEASE use indentation in your code to make it readable. you are using a GOTO to exit your loop and bounce around your programming, that is not good practice ... Finished = 0 READSEQ FileLine FROM FileVar ELSE Finished = 1 LOO...