问题
I have a list of images that I am currently looping through.
Is it possible to have a command where I can put in the location of text that I am trying to extract or show it with something like Citrix Scrape on what data I want to take out?
回答1:
The Get OCR Text action requires you to provide a UI element. One simple option is to loop over all images, then open them in an image viewer of your choice, and then use the Text -> Scrape Relative to extract your invoice number into a variable.
You can use the Desktop Recorder for all actions mentioned above. Note that you will have to adjust the Selectors, as the Recorder usually considers the text the application title as well (which will often include the document's name, and that's not what you want as all your invoices will have different names). You can just open another invoice in your application, then open the Selector and Attach to Live Element to let UiPath work out the Selector for you - which works just fine in most cases.
Depicted below: a sample sequence created with the Desktop Recorder. Here I just read the invoice number with OCR, and I use IrfanView to display them. Of course you may want to increase the zoom level to enhance your OCR results in a real-world scenario.
来源：https://stackoverflow.com/questions/49998081/uipath-rpa-scrape-data-from-images