Issuuing abort after 50 logs

[ameyvaidya]

• 1. What is the Data source? is it a sequential file?
  2. What stage type throws the warnings? is is a transformer? (Recommend pasting a subset of the log.)
  3. what is the difference between the dev env and the testig env WRT the source data. Is there a possibility that the dev env may not have had null values?

Amey.

[kumar_s]

Hi,

1. What is the Data source? is it a sequential file?
yes sequential file
2.What stage type throws the warnings? is is a transformer? (Recommend pasting a subset of the log.)
yes transformer, that is a fatal error with single line
3. what is the difference between the dev env and the testig env WRT the source data. Is there a possibility that the dev env may not have had null values?
No both r same

regards
kumar

[ameyvaidya]

Hi Kumar,
First of all, I apologize for the delay in responding.

two points i've faced before with sequential files and transformers.


1. Setting null values in sequential files for fixed field length data types is very painful. Especially if those files are to be used as staging areas between jobs. The Stage level value(defined in the format page) sometimes does not work. Although this may not be directly relatedto your issue, you can check( by using preview data) in both dev and testing environment to see if metadata defined in both the dev and the testing env's interpret null values properly.

2. When functions are used on nullable columns in transformer derivations or stage variables, Null values need to be specially handled. (Direct move of Null values works.) Transformer functions do not like nulls as input arguments. They tend to do implicit conversions/defaults for nulls and hence the warnings. If you want to use functions with nulable columns, the preferred way is to explicitly convert nulls to a default value before passsing to functions:

eg: numeric testing:

derivation that will cause null value warnings:

Code: Select all

if num(in.col1)=0 then "False" else "True"

derivation that will not cause null value warnings:

Code: Select all

if num(NullToEmptyString(in.col1))=0 then "False" else "True"

(no datastage access so cannot verify the actual return type for num)

Hope that helps.

Amey Vaidya

[kumar_s]

Thank u all,
Combination of Modify stage, NullToValue, Defalut, Nullable feild value helpd.

warning have reduced to very limited, and that to only to the limiation of sequential stage usage in the parallel edition.

I hope that warining occurs often "Sequential file Cannot preserve partition"

regards
kumar