Hi Patrick,

Internally the representation of collections and documents is almost identical. All documents of a collection are stored in a single table. With respect to updates there should be no big difference.

The only benefit of a collection of documents is the direct access to the documents by their name. If you don't access the documents by name but by their molecule-id attribute then one big document might be an alternative.

Jan

On Jun 17, 2010, at 10:34, Patrick Schäfer wrote:

Hi Jan,

thank you for your response. I thought that all documents within a collection are stored within one (or several) relational tables?

I am not sure, if merging all documents into one big document is an option. The documents are supposed to be updateable ... does that give a good performance for one large document?

Patrick

...

Message: 7 Date: Tue, 15 Jun 2010 21:58:44 +0200 From: Jan Rittinger jan.rittinger@uni-tuebingen.de Subject: Re: [MonetDB-users] MonetDB XML - slow execution of query with test on attribute To: Communication channel for MonetDB users monetdb-users@lists.sourceforge.net Message-ID: 702CE627-774E-47C5-B3B7-34A5431F20CF@uni-tuebingen.de Content-Type: text/plain; charset="windows-1252"

Hi Patrick,

the evaluation plan of your query looks quite bad -- with respect to what could be possible...

If you could use 'doc("tu")' (with a single big document containing all small documents), you would get the result in *no* time, as then the attribute index is used. Unfortunately I'm not aware of the difference in the attribute index with respect to collections...

Regards, Jan

---

ThinkGeek and WIRED's GeekDad team up for the Ultimate GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the lucky parental unit. See the prize list and enter to win: http://p.sf.net/sfu/thinkgeek-promo_________________________________________... MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users