Fast & Simple Data Extraction From XML

[laura p]

Hi,

I have a substantial xml file, which represents a large chunk of system
data. In one part of my application I want to retrieve three or four
elements from this xml.

At this point, I do not have the XML stored in memory as a DOM. It
seems like a lot of processing to convert it into a DOM just to
retrieve a few elements, likewise, I have no idea where in the file the
elements are, so it take some time to parse it using SAX.

Does anyone have any suggestions on how I could simply and quickly
extract the elements I need without having to process the whole xml
file?

Thank you,
Laura

[Ted Dunning]

Re: Fast & Simple Data Extraction From XML


If you are using SAX to parse the file, you can use the event based
model to find the elements you are interested in and then abort
processing the file.

Even cooler, however, would be to essentially store an index of file
offsets to various sub-elements in a threaded persistent data
structure. That would let you extract bits (after paying the indexing
cost) at very high speed.

If you were to say more about your application such as whether you have
to search the file multiple times and whether it changes, then others
would be able to comment more specifically.

I hope that this helps anyway.