split xml into smaller xmls based on a split "tag"

xmlsplitter.py

#!/usr/bin/env python
import argparse
import lxml.etree
import os, sys
import bz2

parser = argparse.ArgumentParser()
parser.add_argument('tag')
parser.add_argument('n',default=100000)
parser.add_argument('wrapper', default=None)
parser.add_argument('basename', default='file')

opts = parser.parse_args()

#if(opts.fname=='-'):
#    f = sys.stdin.encoding('UTF-8')
#else:
#    f = open(opts.fname,'rb', encoding="UTF-8")

splitcount = 0
fileindex = 0
d = b""
with sys.stdin.buffer as f:
    context = lxml.etree.iterparse(f, events=('end', ))
    for event, elem in context:
        if elem.tag == opts.tag:
            splitcount += 1
            d = d + lxml.etree.tostring(elem)
            if((splitcount % int(opts.n))==0):
                splitcount = 0
                filename = format(opts.basename + "-" + str(fileindex) + ".xml.bz2")
                fileindex += 1
                print(fileindex)
                with bz2.BZ2File(filename, 'wb') as f:
                    f.write(b"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n")
                    if(opts.wrapper is not None):                        
                        f.write("<{}>".format(opts.wrapper).encode())
                        f.write(d)
                        f.write("</{}>".format(opts.wrapper).encode())
                    else:
                        f.write(d)
                d = b""
        elem.clear()

This code results in every tag having empty contents.