Saurabh

• Feb 23rd, 2013

 

I am assuming the source ia a flat file with no header record but footer record for count. You can have a set of unix command in a unix script and execute the script from pre-session command. Check the Fail task if any tasks fails in the properties tab of pre-session command.

Data in test_file.txt
Record1
Record2
Record3
Record4
Record5
5

Script:
record_count=`sed /^$/d test_file.txt | sed $d | wc -l`
footer_count=`sed /^$/d test_file.txt | sed -n $p`
if [[ $record_count -eq $footer_count ]];
then exit 0
else exit 1
fi

saurav

• Apr 3rd, 2013

 

You can utilize the MD5 hash functionality within the mapping

vijay

• Jan 6th, 2014

 

Design a mapping with 2 flow.1 to load the data into target with all necessary columns.2nd target will have only count information and set a target load plan as said above.In the session level. Set the output files to have both targets with same name.And in the properties set the option to merge/append the files if they alreay exists.This would work out..