DSXchange

There is one user defined Merge routine is present in my job. When i'm running the job at the this point(Merge routine) it's throwing an error as "nls_map_buffer_out() - NLS". I think there is no space in the buffer, i've to clear the buffer. If this is the reason then could anyone pls tel...

Thanks for the help

I'm extracting data from .csv file and putting it into a hash file.
I'm just doing direct mapping.I've input key column values as 1 to 100.

But in output the value of key column are not in the same order as that of input.

Pls help me out..




Thanks in advance

Can anyone tell me what's the difference between hash "ed" file and "hash" file?

Thanks ray..

I have one input file around 2000 fixed width columns.I have to search text from this columns. Directly i can't search from multiple columns, so i thought of merging all columns into one single column.This i'm doing by using a ":" operator in stage variable. E.g: col1(5) col2(2) col3(3) St...

I think you're calling "DSLogFatal" function.This function never return to the before/after sub routine.It's better if you used "DSLogWarn".

First you can sort the columns in ASC giving priority to 1st col and then 2nd. After sorting consider your target file as Hash lookup. Output.Col1=AfterSort.Row_Id Output.Col2=VarRowTemp Where VarRowTemp is :- "If VarRowPrev <> VarRowCurr Then AfterSort.Row_Val Else "N"" VarRowPr...

You can use "Left" function or "Substring".

If you're not sure of the length of the date format i,e. DD/MM/YYYY(10 Characters) or YYYYMMDD(8 Characters) then you can do like this:


Output:- Left(DSLink8.TIMESTAMP,Index(DSLink8.TIMESTAMP,' ',1))

Sorry the previous reply had some junk characters.

You can do like this:-

Output.Col1=Input.Col1[1,3]
Output.Col2=Input.Col1[4,len(Col1)]
Output.Col3=Input.Col2

You can do like this:-

Output.Col1=Input.Col1[1,3]
Output.Col2=Input.Col1[4,len(Col1)]
Output.Col3=Input.Col2[/quote][/list]

Thanks ray for suggestion.
I found one solution for copying column names.
If we create one sequential file job and "include column name as header" we'll get all the columns names.

You can also do by using a hash look up and using the same lookup job as reference in the same job itself.

I'm using DS server 7.1 version on Unix environment. I want to append multiple coulmn names into a single. Say for e.g: Input Col Output Col a e b c d where the value of e=a:b:c:d Since in Input have nearly 2000 columns it's a daunting task to append the column names. Can anybody suggest me a simple...

You can do by selecting the option "Clear the table and then Insert rows" in DB2 stage