a pdf is great for viewing but not for using the data inside of it since a pdf is simply a replacment for a piece of paper. You need some way of selecting an area and then using optical character recognition to get the data from that area and put it into machine readable format
One such solution is https://docparser.com/features
As you see the specification is very vague so first of all you have to answer/decide about the platforms used. Then create a short prototype w/o filtering and then you may think about the filtering. If you decide about the platform then search Internet which is full of code samples. Simply ask for e.g. "Download PDF to C# from SQL Server code sample" etc.