DSXchange

I have a parallel job that just reads a PX dataset and writes to a Teradata table using the Teradata Enterprise Stage (aka terawrite). It writes using the replace mode. I got it to work a couple of times, and now it won't work at all. The program has not changed. It would work once, then fail a few ...

Can anyone tell me mroe about the APT_TERA_64K_BUFFERS (boolean) variable? The description says 'Enables the use of 64K buffers for the terawrite operator." The documentation says "Orchestrate assumes that the terawrite operator writes to buffers whose maximum size is 32 KB. Enable the use...

You can use the before/after sections of the Multiload stage. Syntax depends on if you have named indexes or not. I'll use both types in the example: In the before section: drop unique index (field1, field2) on mydb.mytable; -- or -- drop unique index myindxname on mydb.mytable; In the after section...

I guess You're trying to load a terradata tables isn't it ?? If so did you set "-sessionsperplayer" and "-requestedsessions " as DB options ??? All the best I copied this from a posting by l_homme_de_rio made in response to "multiload utility in Parallel mode", http://...

We use DB2 with PX as well. My understanding is that a true load (i.e. uses DB2's autoloader) cannot be rolled back. If a load fails, you can get the table released (from the load pending state) and the table will contain any data that had successfully loaded, but would not roll back to the original...

If you have exported the DSX to a file, you can pull the schemas from that. Open the DSX in an editor and search for the word schema. It should look something like this: -schema record {record_length=fixed, delim=none, ebcdic, native_endian, binary} ( FIB_CDC_APPLICATION_ID:string[15]; FIB_CDC_PAGE_...

If you are using custom buildops, you can take advantage of some of the C/C++ built-in functions. We wrote a function called stringToGraph that takes a string as input, loops through each character and uses the C macro isgraph(c) to test if character c is a graphic character (letter, number, type of...

Check your import schema. The messages you are getting are the same as those you would get if there was a problem with your Sequential File Stage import during processing. The browser has to import the data, too. In your schema, make sure you define the format of the data coming in (in the Edit Colu...

We use Surrogate Keys a lot. You can use a variable (job parameter) to seed the generated id or always start with 1 and in the same job pull the max id + 1 of your target table (into which your new data will be loaded). Then in a transformer, as mentioned by T42, you can add the generated it to the ...

You can use integer, but you must use an unsigned integer. A signed integer can hold values from -2 ^ 31 through +(2^31) - 1, or -2147483648 to +2147483647 (approx -2.1 billion to +2.1 billion) An unsigned integer can hold from 0 to +(2^32) - 1 = +4294967295 (0 to approx. +4.3 billion) If you know t...

What criteria did you use for your surrogate key? What data is going into it? I am wondering if the error is actually related to the data going into the generator as opposed to what the generator is doing. The Generator can only create a numeric output field and doesn't do anything with timestamp. A...

We have into this as well. Here's what we have done: 1) After each import (on the output link, that is), add a Generator stage to add a file_id field that is hardcoded to 1 or 2 (depending on which link it is attached to) 2) Funnel the results together 3) Sort on your data keys (billing_nr, ship_nr,...

Yes, it helps to have some C/C++ background, but most of our staff had little or no background when they started. All it takes is one or two people that DO know C and/or C++ and the ball starts rolling. I just don't want people to get scared off when they see "C/C++". For the most part, it...

Just wanted to throw out another option, since I haven't seen it mentioend yet, here or in other related topics. We use Buildops at our shop. They handle the ETL all at once in one buildop stage rather than a multitude of other stages. A lot of our data comes from source that use occurences. Mostly ...

Remember, in PX, you need to think 'parallel'. Once a sequential file is read into a dataset, the dataset is stored so that PX can access it in parallel. This is great when multiple processes may need access to the same sequential file. Rather than have multiple processes read from the file, have on...