Good morning Gabriel,
To be sure that I have understood the input files correctly.
Each file is named for the code that you need to analyse. Are the files the same format so that they can use the same model to extract the data fields you need?
If so, is it still possible that they also include data fields that will be trapped and extracted by the model but that you do not need for the data to be extracted from that report?
Or is it that the report format is always the same but the only way you know which code you are reporting about is by the report name?
If it is this last situation you can make the Report Name a field that is extracted and then reduce the name to just the part of it that you need using a calculated field (Formula Field).
Google translate suggests a slightly different nuanced description of the problem to the one you present in your English version. When I try to relate them to how Monarch works and I am not sure which of them best fits the challenge you need to resolve.
Thank for the answer.
The problem is:
I have a report to make in output.
But we want to use only one process that will have in input: a manifest (don't know how to extract data from it (I tried the doc) and a data file that will contain the data to analyze.
In the manifest, we will get a modelname, and from this modelname, we will generate the report from the data file by applying this model.
Each manifest will have a different modelname, and as such, we will use differents models to apply to the data file.
And we want it in one process.
So my question is how to have differents models selected in one process by a manifest.
You have mentioned "manifest" which is an Server (Automator) related feature.
Server (Automator) is a product very much related to Monarch but has its own Forum section in the Community.
If my understanding of this is correct you are likely to get a better response to the use of manifests from those people who watch the Server (Automator) forum. So you might benefit from posting the question over there. There has been some recent discussion about manifests in Server in that forum.
I have some knowledge of Server (Automator) but not specifically the manifest capabilities.
If I have misunderstood your need (it does read very much like a Server activity) then I apologise.
Since you have version 14 it is likely that you also have maintenance and technical support. Have you already asked for assistance from Datawatch?
Based on the manifest xsd the manifest xml file can have multiple <ItemType>ModelFile</ItemType> elements.
You have two options, 1) create Visual process that process a single input, model, and export, and pass multiple XML manifest. Each manifest xml file will have the same input reference, but different model reference, and possibly the same or different export reference.
2) Create a Visual Process with multiple model objects, then create an xml manifest file with multiple <ItemType>ModelFile</ItemType> elements for each model object in the Visual Process.
I have not tested the second option, but will add it to do list.
Do you still have this requirement?
If you are have Automator (the server product), then you can define a project that applies a specific model based on your file name. You can have many projects in a single process. It sounds like the Server product will solve your problem.
If you don't have the server product, then I'm afraid it is not possible to automatically swap the model depending on the file name or any other aspect of the report.
What is the reason for needing a single process? Maybe there is another way around this problem?