Gathering job statistics could help you to detect proactively problems in your data by comparing current values with their estimated values. The idea behind is to collect as much as possible information on your daily run processes (number of records, times it take, number of missing (etc.)) and store them in an historical table. If the values of the current run differs with the values of previous run for the same jobs, this could indicate potential problem.

In this tips and tricks, we will first have a look to all the information we could collect on our jobs. This could be for example:

• Number of records
• Number of missing values
• Number of duplicates
• Number of event
• Execution time
• (etc.)

Secondly, we will discuss how we could collect this information without influencing the performance of the current jobs and thirdly, we will create a reusable macro program that extract this information and store it into a table. Guess what will be the next step? If the value of a given parameter is varying to much from the normality (ex: 2 std) this means that something unexpected probably occurred (etc.).

You have some SAS experience and would like to automate your SAS code execution.