How do I pass in, or build, a file spec that includes a message property?
If I could enter a JSONata Expression, it would be something like:
msg.SearchMask & “*.pdf”
Are there plans to add the full range of “types” to the File Spec? Path, msg properties, msg.payload, JSONata expression, String, etc.?
Unfortunately this is not possible but it is an interesting suggestion. So far we found people using static file masks using star-based filters or regular expressions. The following is a regex sample from the embedded help:
To find files with a name similar to 294_gm_M10-01_1234_1_BATCH.csv, the following expression could be used:
[0-9]{3}_.*_M.*_.*\.csv
Can you elaborate on your use case?
Yes. I have created a web dashboard for our customer service reps to search for and view a PDF. We produce lots of different statement types so the PDFs are in various folders per type. These PDFs are eventually ingested by a 3rd-party process where customers can view their statements, but there’s a 2-3 day delay and people panic and call in (or they cannot find their statement from X months ago).
The “landing page” is a static HTML file. The “Search Results” page is a list of PDFs that match the folder (statement type) and search mask entered on the landing page.
This is currently achieved with a Folder Listing plugin in Workflow, and a datamapper that directly consumes the XML produced by the Folder Listing. The Template produces a pretty list of PDF files, clicking one opens that PDF in the browser.
What I’ve worked out in Automate so far is that, if I use a change node to set “msg.masks” to what I want, and leave the “File Spec” blank in the Folder Listing node, that works. The folder listing claims there’s an “error” because the masks property isn’t set, but once deployed, it accepts the msg.masks set by the change node. The “folder” and “searchmask” values that are in the XML, I’ll pass in as Parameters instead, and I’ll redesign the Data Mapper to use the output from the folding listing node… well, to use the ManagedFileId property that will result from writing that output to the filestore first.
EDIT: rewriting the data mapper is a challenge, because the folder listing node produces an array of values, not a JSON object, so I’ll have to figure out how to rewrite the results into something Connect can use.
EDIT2: I have managed to create, via a function node, a JSON object that I can use in the data mapper. However, the folder listing node only emits the filenames, it is missing the “created” timestamp, and the “size” property that the Folder Listing plugin in Workflow produces. I need the “time”. My reps search for a loan number, and get back a list of all statements that match that loan number. Usually someone is asking for a missing statement, and that could be from months ago. So I list the date the PDF was created next to the link to the PDF. The “folder listing” node is missing critical functionality!
This is a case where the Folder Listing plugin in Workflow and the Folder Listing node in Automate behave very differently. I don’t mind recreating Automate Flows to replace Workflow Processes, but when I have to re-design Data Maps in order to use Automate, I’ll balk.