The contains() is a linear search and doesn't use indexes. There was talk of adding in a search for all the text. I haven't tried the CDATA searches yet...
hth, Mark "Vorbau, Alex" wrote: > Hi, > > I'm testing out Xindice to see if it meets the needs of our project. I > loaded it with 3,000 XML files and tried some XPATH queries. I've noticed a > couple problems, so I'm hoping that I'm just not using it correctly. > > * a substring search ( i.e. contains() ) results in a linear search which > takes about 40-60 sec, regardless of any indexers that are set. I saw a > reference to this problem in an earlier message, but I wondering if this > will be addressed soon. > > * more importantly, an XPATH query of the exact contents of an element that > is wrapped in CDATA returns nothing. > > For example if there's an element like this: > <foo> > <bar><![CDATA[Bert & Ernie]]></bar> > </foo> > > The XPATH query > //foo[bar = 'Bert & Ernie']" > returns nothing. So does this: > //foo[bar = '<![CDATA[Bert & Ernie]]>']" > > Could this be true? I hope I'm missing something because non-support for > CDATA is definitely a showstopper for us. > > TIA for you help, > > Alex > > ----- > Alex Vorbau > HP Labs, Nomadic Computing Dept. > 650-857-3050 > [EMAIL PROTECTED]
