DSXchange

The convert, not the trim, is removing the letters for you. Does it cause any potential problems? You'd have to tell us - it only removes upper-case letters from the string, is that enough to solve your particular problem? It works in my case actually. And also I use Convert(field, "1234567890...

I don't want to go to Routine. And I use

trim(convert(field1, "ABCDEFG...Z","")) to get rid of letters.

It works in my case, does it cause any potential problems

Somebody can tell me how to get numbers and letters from such a string by trim function.

KBN12390 -> to get KBN and 12390
234BN -> to get 234 and BN.

Thanks

It goes weired. and I add len(trim(field))=0. From my first job, read input data from merge stage, then output to a transformer, transformer outut data to a sequencial file. apply the rules. it looks like some of them work, and some of them don't work well. Then I create another job, read from seque...

I need to put a '#' instead of empty value for all my output fields. then I use if ( trim(Body.UTOADAD)='' Or IsNull(Body.UTOADAD)) then "#" else Body.UTOADAD in transformer. And actually in the output file, there are still some empty value ( at least invisible ) in the output column ? Wha...

That definitely work.

thanks

Not very sure what you mean ?

Even if funnel = sort + funnel, it couldn't cause a bug..

Anyway,how to fix it up..

I am not sure whether it is a very stupid question. Basically, I have a job with one parameter ( of course with default value). There are 3 different stages in my job deal with different tasks. Can I change the value of my job parameter in the first stage, and then my second stage can use the job pa...

use cat file1 file2 file3 > outputfile is fine. But based on my understanding, funnel stage should work if we set the property as sequence, it will read the data from the first input link firstly, then next, then next. But actually, it didn't . Is there any problems when funnel stage output the file...

Actually, the output is only one row. It is the header of output file. After the file generated, I will use cat to merge 3 files into one, they are header, body and tail. The sequence number is a field of the header file. for example if we run out jobs weekly. And the first week is 0001, second week...

Nothing wrong with the filename, there are a few dataset files under the directory. If I choose one of them, it is OK. And if i choose another one, the file name is xxx_xxx_xxxxxxxxx.ds, it comes up with an error. "Unknown error reading record schema". The interesting thing is my next job ...

I try to have a look at the data in my dataset. And I go to designer, and use Tools - Dataset management, and system give me an unknown error.. somebody knows what is wrong ? I am on Linux. but actually my next job and read the dataset file successfully. Is there any name rules for the dataset file,...

It is like a job sequence number, once I kick-off the job once, it was added by 1. Like first time 0001, next time 0002. About where to store, it is a problem. Since for server version, it was stored in a hash file somewhere in the server. For parallel version, I try to use a sequencial file to stor...

Thanks,

After I use DecimalToString(DSLink233.seqNum, "fix_zero,suppress_zero") and it works.

I am not quite sure what "fix_zero,suppress_zero" means

I got an integer number, like 12, And I need to convert it into 0012 char(4). I use DecimalToString(DSLink231.seqNum,'fix_zero'), And PadStringDecimalToString(DSLink231.seqNum,'fix_zero') ,'0', 4) It doesn't work. Then I try Str('0', (4-Len(DecimalToString(DSLink231.seqNum,'fix_zero')))) : DecimalTo...