DSXchange

I begged for too long for a meta-data stage that would enforce the metadata and perform extra validations and such from a meta-library. The silence was deafening. Imagine the ability to "scrub" your data, perform all trims, roundings, scalings, range/domain/pattern checks, etc.

This is going to be executed for every single row going thru the Transformer. It would be better to use it in the initialize step of a stage variable so it's done only once. For the derivation of the stage variable, just use the stage variable.

It's worked this way since Server 5.0 came out. Somewhere around then Ascential (then under Informix banner) engineering decided that metadata does matter. The Sequential file stage will enforce the scale, whereas the Hashed stage doesn't care. This became frightening because soooo many jobs used ju...

If the DS server is already upgraded to 7.5, then you must use the 7.5 client to open the jobs. If they upgraded the server in place with existing projects with jobs, then the jobs "are" upgraded. Whether they will work is a different story. If the DS server is version 6.0 and you need to copy them ...

Welcome aboard. Next time, please post your question in the appropriate forum. We don't know if you're asking about RTI, PX, Server, etc. Either write a function with a series of CASE statements, or have a conversion table constructed and load into a reference source that you can use in a transforme...

In the meantime, put the copy command into a bat script. Have the script keep copying the file until the byte count on the farside matches the byte count on the DS Server. That way, no matter the error, it will sit there and make sure the file gets over. Something like this pseudo code: remove remot...

Oracle's VARCHAR2(4000) is the maximum, so you're limiting yourself to the first 4000 characters selecting and inserting. It's BLOB, so that means you're corrupting your data, since BLOB means BINARY LARGE OBJECT, the first 4000 characters of binary data is useless.

Proceed at your own risk.

I do believe you can't move BLOBS using the Oracle OCI and regular DML. But I could be wrong.

Okay, just looked over at that thread. I originally ignored it because "removing duplicates" is just about the most hated topic for me. Hashed files are great for 1:1 reference purposes. While there is fantastic functionality inherent in DS BASIC for working with them, using DataStage and UV/SQL is ...

If you're using the UV/ODBC stage as a source stage, use the primary inputs section. Filenames are case sensitive, so makesure that's okay. Keep in mind this is not going to be fast, hashed files are not setup with secondary indexes to optimize relational type joins. If your volumes allow, consider ...

Your Routine seems to be saying: Error Code from the copy command is 1, and the Screen Output was "The semaphore timeout period has expired. " This is the operating system, not DataStage. You shouldn't write a job to do a file copy, the OS should do it way better and faster. Your issue is related to...

Best and fastest are subjective. I can't believe your archiving solution isn't tied to the delete solution. If that's true, you're talking about full scan the live table and compare with the archive table, and where matched you'll delete from the archive table. The fastest solution if these two tabl...

If anyone cares, the original problem was MULTIPLE CR/LF at the END of the string, so you need to remove as many as you can from the end, preserving the rest of the formatting. My original answer still looks to be the only one that solves that issue.

I like to analogize Multi-Value by saying you store arrays instead of values. Being an aerospace engineer by education, we did a lot of matrix math, so I was quite comfortable with the idea of thinking in 3-D matrix form for numbers, now it's just dates and quantities and dollars. Actually, I did ch...

Multi-value fields (attributes) are simply strings, so they can be quite large. The entire row (record) can be quite large, megabytes if you want. The problem with large numbers of multivalues in an attribute is just retrieval time, but don't think that's a big deal in this case. Keep in mind in the...