I am looking for some information when comparing these two ways to process:
I Daily run my robots using Excel file as a processfile.
Excel basicly consists of 1 column with an unique ID. (Several other columns g’ets filled with data From systems)
This ID is used in several systems. All robot actions Will be logged in the same Excel file.
Running a dispatcher that uses the excel file to upload ID’s as queue item and then using a performer to do the same processes as in 1) .
Which method of processing is “best practice” or is best being used?
I hope it makes sense - otherwise feel Free to ask
Hi
I would got for second option
Because
—when multiple systems accessing a same file first there might be a chance of file getting corrupted as the file would become read only might take some time to fill in new data until the previous data is getting saved
So when using queue we won’t have that problem as Queue helps us to manage data individually as a item especially while processing and updating
—more over sometime due to system exception the data might not get processed and if with first option then those data might be missed to get processed
Or get updated with no other corresponding records
While in queue we can get to know which got transacted and which failed so that they can be even reviewed and retried once again
—and finally as we are separating the process of transaction from excel to queue to web
With dispatcher and performer the allocation of bot would be easier especially while scheduling them
Even we can include START JOB activity at the end in the dispatcher process itself so that once dispatcher is dine this START JOB activity will call the Performar process and continue thereby