couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Antony Blakey <antony.bla...@gmail.com>
Subject 'prefixkey' query parameter
Date Sat, 31 Jan 2009 22:57:55 GMT
Hi,
    every time I see someone talking about \u9999 in a key string I'm  
reminded that it isn't really correct. On top of the trivial issue of  
\u9999 being the wrong value, using a unicode escape like this is  
problematic because Unicode is not a 16 bit character set, and as  
Unicode evolves the character set increases.

I think a better solution is to add a prefixkey=<xx> query parameter,  
that does the equivalent of the current \uXXXX and {} techniques using  
startkey and endkey, extended for all JSON types.

----------------------------------------------------

'abc' would match all strings starting with abc

numbers and true/false/null would match exactly
{ ... } would match exactly (because there is no ordering in hashes)

[ a, b ] is the interesting case. It would match recursively, with the  
last element of the list recursively using prefix matching

e.g.:

[ 'abc', 2 ] matches all lists with [ 'abc', 2 ] as a head
[ 2 , 'abc' ] matches all lists with 2 as the first element, and any  
string starting with 'abc' as the second element, and an arbitrary tail.

----------------------------------------------------

Is there any support for this?

One alternative would be to allow the specification of the key test  
functions as a query parameter, which would enable inclusive/exclusive/ 
prefix ranges

e.g. (ignoring the syntax and names)

start-key=greater
startkey-test=greater-or-equals
startkey-test=prefix

and the same for endkey-test. This is a more complete solution.

Antony Blakey
--------------------------
CTO, Linkuistics Pty Ltd
Ph: 0438 840 787

Isn't it enough to see that a garden is beautiful without having to  
believe that there are fairies at the bottom of it too?
   -- Douglas Adams


Mime
View raw message