Hi!
I am trying to create a PDF parser for project reports (structured), and I want the extracted data (both text, numbers and table entries) from every report to be added on a single row in an excel spreadsheet, as a reporting overview for our CFO. In other words one row per project, with all the relevant data on that row. I successfully did this a year ago by only looking for 'Text fields' in ai builder, and I got everything working in power automate, but the ai model had low accuracy.
When I now try to change the ai model to read numbers as numbers, and the tables as tables instead of just text fields, the accuracy of the model is a lot better (naturally). But I don't understand how I can access the individual table entries in my flow in order to copy them to a specific column in my spreadsheet..? All the other extracted data is available as dynamic content, but the table entries are not. Do I need some intermediate step to "unpack" the tables? I know that the ai model correctly extracts the tables, I just don't know how to access the specific table entries in an "add/update a row in a table" action.
Thanks in advance
I have found that I can initialize an array variable, and append the values from a specific extracted table column into this array. Then I can use the array index# to copy the individual cell entries. However, I don't think this is a reliable method, as it returns an error if I have set it up to copy a certain amount of entries. If one of the table cells are empty, there will be one less entry, and the flow will fail.
Surely there is a better way to achieve this? my understanding of PA and Ai Builder is quite basic, so apologies in advance if I am asking dumb questions