DSXchange

Hi,

I am trying to validate a sting that comes in from a file. I am taking a set of characters and I need to make sure that they meet a certain criteria. The character string will be in a fixed position in a file, 3500-3510 to be exact.

First off, I need to make sure that there are no spaces and zeros. Then I need to make sure that the positions are meeting certain requirements, for example:

11 characters
char 1 has to be number 1-9
char 2 has to be A-Z
char 3 has to be 0-9 and A-Z
char 4 has to be 0-9
char 5 has to be A-Z
char 6 has to be 0-9 and A-Z
char 7 has to be 0-9
char 8 has to be A-Z
char 9 has to be A-Z
char 10 has to be 0-9
char 11 has to be 0-9

Thanks

Sounds like a job for a 'regular expression', something the Filter stage supports as one example.

Im using this in the transformer using stage variables to create an error code. I figured it would be an If statement but I didnt know what the actual working of that would look like

Much depends on what you do next when one of your tests is failed by the data being examined.

The simplest approach is to do each test separately. For example (pseudo-code expression): This allows you to prioritize the tests. For example, if there are spaces or zeroes in the full string, you might want to skip the other tests because this is a "fatal" error for the data.

Appendix B in the Parallel Job Developer's Guide is excellent reading, especially the String section.

You might find this post help you:
viewtopic.php?t=107882

FranklinE,

How would I use the Num function to get rid of the 0? For ex: Num(Link[3500,1]) will see if the 1st char is 0-9. Is there a manipulation that will be able to make sure that 0 is not included?

Thanks

Not via the NUM function, you would need to explicitly check your valid values or valid range instead.

So for somebody with no DB2 knowledge, I could use ISNUMERIC for the 1st character. Is there a ISALPHA and an ISALPHANUM function as well. I am referring to your 'regular expression' comment from earlier.

if it becomes too complicated to do in a transformer with string functions and you don't want to fight regular expressions you can also write a basic routine (or if you want to go external, any language) to do it as a third option.

You may also be able to use the IA/rules stage (?) for it. I have not used that much.

Ditto to what UCDI posted, but there's an implied issue here that seems odd to me.

I get that transformer stages have large footprints. I get that there may be more efficient ways of doing things if one has extended knowledge of those other methods (BASIC, etc.). What I don't get is that the entire point of a rapid application development tool is that one uses out-of-the-box methods as much as possible.

I have limited ability with BASIC (last used it extensively in high school, in the dark ages of phone connection to a time share), so grain of salt: if the simplest and most direct approach is using a transformer stage and String transform functions, why would another approach be considered here?

Scott, don't get me wrong. If you see value in one of the alternatives, that's your best choice. I suggest at least researching the String functions, perhaps with an eye to finding their equivalences in the other approach(es).

I think what it comes down to is that 'rapid' is subjective. I can write a C++ or basic routine to do advanced string processing in a couple of min and have it debugged and delivered inside an hour (problems on par with the question posed here). Trying to build that in datastage is the round hole and square peg problem... you can do it, but it is overly convoluted because the DS string processing routines, while they cover 99% of what one needs to do, are very simplistic and crude do when you hit a complex problem in the 1% it becomes a valid choice *for those who know the low level languages* to do it another way. If you are uncomfortable with the low level coding, that approach may not be 'rapid' or viable for you. I happen to have spent 15+ years writing C++. I may as well make use of that from time to time

DataStage 11.5 built-in parallel string functions are documented here:

https://www.ibm.com/support/knowledgece ... tions.html

Substring Operator is documented here (one of several places):

https://www.ibm.com/support/knowledgece ... rator.html