vens28

 Profile Answers by vens28 

• Apr 28th, 2009

 

Error handling Logic we are using expression transformation to find the data is coming in expected dormat  or not suppose date range dd/mm/yyyy from source data loaded like mm/dd/yyyy then its not matching our requirement for the particular date port then its send error message to Error handling target

Sumithav29

 Profile Answers by Sumithav29 

• Jun 1st, 2009

 

After validating the incoming source data, if we find any invalid data(eg: data not present the dimension table, null data, or difference in the data format) then it can be loaded to a reject table(include one remark column which will help us to identify the cause of rejection). Rerun the workflow after correcting the data in source based on the cause of rejection.

sahan

 Profile Answers by sahan 

• Jul 16th, 2009

 

Bad files contains column indicator and row indicator.
Row indicator: It generally happens when working with update strategy transformation. The writer/target rejects the rows going to the target

Columnindicator:
D -valid
o - overflow
n - null
t - truncate
When the data is with nulls, or overflow it will be rejected to write the data to the target
The reject data is stored on reject files. You can check the data and reload the data in to the target using reject reload utility.

prashant.baranwal

 Profile Answers by prashant.baranwal 

• Sep 11th, 2010

 

It is one of the most common concept which is being implement in alomost all DW projects. A very simple way of handing error and with their approprite messages is:

Concat all the columns with a delimeter and say column "Record" contains all the column concatenated.
Check for errors for each row and assign an error message for such rows. Suppose you are handling a total of 7 type of errors. Errors could be, NULL date filed, Empty Date of birth. Start Date is greatre than End Date, empty account number etc..
Now use a normalizer to create errors of each of the row.
 After that filterout all those rows which have empty error messages.
Now you can store the records in a table. Rectify them and pust them again in source file next time or develop a mechanism to restore them.

Umesh

• May 14th, 2012

 

So we create 2 variable for every field on which we have to do error logging. 1 variable will check the condition for error (i.e. check for isnull or check for datatype) and assign NULL if it does not catch error, 2nd variable will be used for assigning a message for every error(e.g Name is invalid, date is invalid). We will concatenate all the variable for the checks and if the output is Null that means there is no error otherwise there is an error. To identify which field is having error we will concatenate the message on the 2nd variable .