In the sample below there are two different approaches described how you can start jobs on a file event.
- Using file order source
the job chain job_chain_file_order_source has a file order source looking for files in c:\temp\input ending with .txt. When a file arrives, an order for the job chain will be created. The file will be erased at the end by the file sinks.
- Orders can not have parameters
- Order starts when file arrives, not when file is completly written
- Using JITL Job JobSchedulerExistsFile
the job chain job_chain_file_order_source_poll uses the job JobSchedulerExistsFile to check, wether a file is in c:\temp\input ending with .pdf. If so, for each file found, an order will be created for the job chain job_chain_file_order_source. The file be erased by the sinks in the executed job chain.
- The created order have the same parameters as the polling order
- The first step in the executed job chain checks, wether the file is completly written.
Please note, that you can mix file order source and JobSchedulerExistsFile by using JobSchedulerExistsFile in the first node to check, whether the file is completly written.