I have been working on a flow to extract metadata from a structured document through AI Builder using the "Recognize text in an image" feature on MS Power Automate cloud.
The document has multiple pages (30-35) and I wish to extract a specific word from that.
For Eg. The required word is "Resistance"
Now, the issue is, "Resistance" occurs randomly on multiple pages, multiple times i.e it can occur 3-4 time on one page and never on another.
I am able to extract the index of the word successfully, but since that word has multiple occurrences on multiple pages, the output is the index of that word on each individual page, which varies from page to page i.e. ( 49 on page 5, 6 on page 17, no occurence on page 1 etc.)
I want the output to be that particular word, returned to me as text, irrespective of its multiple indexes.
Currently, I am going like: Recognize text in an image-Compose (All pages) ---Parse JSON-Select all text (This creates and Apply to all loop) ---Insert index---Filter array for first occurrence "Resistance"(But this filters out occurrence on each page individually) and so on.
How can I approach this?
