Why is xml.dom.minidom so slow?

[Bjorn Pettersen]

Background: I've got 750 accounts in xml format in a 50 Mb file. I can
extract the string for an individual account at a rate of >290
accts/sec. When I add a call to parseString() afterwards however, the
speed drops to 0.75 accts/sec.

All I'm doing boils down to:

  response = rf.nextResponse()
  dom = parseString(response)

in a loop. Am I doing something wrong? Is there a faster way when all I
need is a traversable tree structure as the result?

For our requirements we can't use anything under 10 accts/sec without
impacting our SLAs...

-- bjorn