James,
You're right -- the data is mixed. For example, you might get s2 or
pt2 for an issue or a roman numeral for a volume. The cast fails on
these but the collation based sort seems to treat them as text and
keeps going.
Alan
On 4-Jun-07, at 6:52 PM, James A. Robinson wrote:
I have a set of documents that have volume and issue numbers
represented as strings (e.g. 1, 10, 2, 23). If I use the following
collation, I get a nice numeric sort without having to pad or
otherwise normalize the data, but performance suffers. Is there an
index I can add that would speed up the kind of sorting that this
collation needs to do?
for $i in cts:search(doc()//article-title,"water")
order by $i/root()//volume collation "http://marklogic.com/collation/
en/MO"
return
$i/root()//volume
I was curious re the use of collataion -- are your values actually
numeric
values? If so I'd have assumed a way to do this sort of sorting would
be something like
for $d in $set
order by xs:decimal($d/volume), xs:decimal($d/issue)
return ...
and so forth. Is this naive? I understand why it wouldn't work for
things like '10a' or something, but for pure numbers, even if they are
padded with leading zeros, I had assumed a cast would work.
Jim
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
James A. Robinson [EMAIL PROTECTED]
Stanford University HighWire Press http://highwire.stanford.edu/
+1 650 7237294 (Work) +1 650 7259335 (Fax)
_______________________________________________
General mailing list
[email protected]
http://xqzone.com/mailman/listinfo/general
_______________________________________________
General mailing list
[email protected]
http://xqzone.com/mailman/listinfo/general
