I am using Directory and filepath name read and then FilePropertyExtractor to build a table with all files of a specific file type. I then use time stamper to create the file name. subsequently the workspace is run a second time and the previous file (using sorter and sampler) is pulled into feature reader and then to the ChangeDetector.
I can't seem to add a uuid because it recreates a new one each time i run it. So i can't use that as an update key.
So right now i am using the unix path as an update key and i get all features as updated. (when there are no changes)
If i set the check attr types and encodings to yes i get all of the original as deleted and all the revised as inserted.
I just read that the order makes a big difference in the CDor. So i rearranged a bit and added sorters. Then
I had to reformat my date modified attribute. and i think that might be it.
I would take any suggestions anyone has to make this better. It is going to have ot run over an entire network (which previously took over 3 days to run).
The Matcher doesn't do what i need and i can't find anything else that will help with this.