Extracts application numbers from XML downloads from https://pairbulkdata.uspto.gov

ExtractAppNumbersFromPBD.groovy

if (args.length < 1) {
    println 'Please provide an argument pointing to a bulk data XML file.'
    System.exit(1)
}

import java.nio.file.*
import java.nio.charset.Charset
import javax.xml.stream.*
import javax.xml.namespace.QName

def xml = Files.newBufferedReader(new File(args[0]).toPath(), Charset.forName('UTF-8'))
def r = XMLInputFactory.newInstance().createXMLStreamReader(xml)

boolean isApplicationNumber = false
StringBuilder appNumber = new StringBuilder()
while (r.hasNext()) {
    int eventType = r.next()
    if (eventType == XMLStreamConstants.START_ELEMENT) {
        isApplicationNumber = 'uscom:ApplicationNumberText' == "${r.prefix}:${r.localName}".toString()
    } else if (eventType == XMLStreamConstants.CHARACTERS) {
        appNumber << new String(r.textCharacters, r.textStart, r.textLength)
    } else if (eventType == XMLStreamConstants.END_ELEMENT) {
        if (isApplicationNumber) {
            println appNumber.toString().replaceAll(/\s/, '')
        }
        isApplicationNumber = false
        appNumber = new StringBuilder()
    }
}

Usage: groovy ExtractAppNumbersFromPBD.groovy C:\whatever\2017.xml >> app_nums.txt