DSXchange

Its a basic function and hence will work in a server job or for a basic transformer in px.

The logs that you are retrieving, are they huge? As in, how often do you purge the logs. It takes longer to retrieve and go through 7 days of logs than to go through 2 days of logs. It might just be a shell issue. As you said if you run the dsjob -lparam command from command line its fast but within...

As sud asked, what you wanted to do, we require that answer to help direct your in the right direction, or was it just your curiousity?

...and you could have gotton that answer just by a simple search. A similar post was answered today.

Its hidden. Like other transform functions. Its nothing but a simple env command run during run time. Thats it. Ok maybe a bit more but more than likely, its just that.

A number of ways. You can use sort stage or remove duplicate stage. Referhere.

Sure you can. Have an execute command stage do a cat of the file which has the list in a delimitted format (pre-massage the file if its not in that format). Connect this command stage to start loop and in the list specify #Execute_Command_Stage.$CommandOutput# . Another way would be to build a serve...

Search on keywords like 'dsjob -param'

20 users is not much. It would be fine if you still go and do a select at the database level. But if you want to do aggregations and order by's in your sql select, the dba might have a problem with it. There are groups that advocate both. Its upto you, depending upon your system, what way you want t...

You are right. Do a table unload to the staging table/file/dateset and go from there, especially if your source will be heavily loaded. But then again it really depends upon what your architect desinged as an inhouse standard.

You might be right. Just a thought that occured in my mind. I needed clarification on that too. Thanks. So a connection per database stage it is

Thats fine. YOu can do it this way. If performance starts to bite, put idices on the date fields. If thats still not enough. Load your source and reference data into work tables and perform the logic at the database level.

More line one connection per link going to a database stage. So if you have two links going to a database stage, two connections will be request. Right

A bunch of errors. The list is vast. In general, anything thats against the rules of table definition will be rejected. Like data truncation errors, date format incorrect, fk constraints etc. etc.

No need to slow down your job by using a basic transformer just to use ICONV/OCONV. Read your date as varchar and inside a px transformer use the following derivation. StringToDate(in.Date, "%dd-%mmm-%yyyy") In the output stage, go to the column's extended properties and change the...