Xapian discuss - May 2006 - Phrase Query vs AND Query? Why don't these find the same things?

PHRASE QUERY
Xapian::Query((MBOX:12345678-1234-1234-1234-1234567890ab AND
(LP:backup:(pos=1) PHRASE 6 LP:c::(pos=2) PHRASE 6 LP:program files:(pos=3)
PHRASE 6 LP:Mozilla Firefox:(pos=4) PHRASE 6 LP:res:(pos=5) PHRASE 6 LP:
table-add-column-after-hover.gif:(pos=6))))

AND QUERY
Xapian::Query((MBOX:12345678-1234-1234-1234-1234567890ab AND
LP:backup:(pos=1) AND LP:c::(pos=2) AND LP:program files:(pos=3) AND
LP:Mozilla Firefox:(pos=4) AND LP:res:(pos=5) AND LP:
table-add-column-after-hover.gif:(pos=6)))

given that I have done .add_posting() correctly with the LP: prefixed terms
and the second AND Query returns the set of data I expect, why does the
first query NOT return ANYTHING?

here is what DELVE tells me
>delve -r 1 -1 -d -v -k /index/wfs/
Values for record #1:
1:/backup/c:/program files/Mozilla Firefox/res/table-
add-column-after-hover.gif
Data for record #1:
/wfs/0/+/D/0+DM0wZ6y_mHubbNJjQXrgqW64s=/table-add-column-after-hover.gif
/12345678-1234-1234-1234-1234567890ab
Term List for record #1:
CDATE:1135755971 1 1
ETAG:DA39A3EE5E6B4B0D3255BFEF95601890AFD80709 1 1
LANG:en 1 1
LEN:0 1 1
LP:Mozilla Firefox 1 1
LP:backup 1 1
LP:c: 1 1
LP:program files 1 1
LP:res 1 1
LP:table-add-column-after-hover.gif 1 1
MBOX:12345678-1234-1234-1234-1234567890ab 1 1
MDATE:1135755971 1 1
MIME:application/octet-stream 1 1
NAME:table-add-column-after-hover.gif 1 1
TYPE:0 1 1

here is what my test program outputs with the AND QUERY
the numbers in the [] after the LP: prefixed terms are thier posting
postiions

found 1
DOCID:1
PPATH:/wfs/0/+/D/0+DM0wZ6y_mHubbNJjQXrgqW64s=/table-
add-column-after-hover.gif/12345678-1234-1234-1234-1234567890ab
LPATH:/backup/c:/program files/Mozilla Firefox/res/table-
add-column-after-hover.gif
CDATE:1135755971 (Wed, 28 Dec 2005 07:46:11 GMT)
ETAG:DA39A3EE5E6B4B0D3255BFEF95601890AFD80709
LANG:en
LEN:0
LP:Mozilla Firefox [4]
LP:backup [1]
LP:c: [2]
LP:program files [3]
LP:res [5]
LP:table-add-column-after-hover.gif [6]
MBOX:12345678-1234-1234-1234-1234567890ab
MDATE:1135755971 (Wed, 28 Dec 2005 07:46:11 GMT)
MIME:application/octet-stream
NAME:table-add-column-after-hover.gif
TYPE:0 (file)

anyone have any ideas how to get a PHRASE QUERY to return the data?
also what will ulimate be "better" given the context of these queries,
the
AND QUERY or the PHRASE QUERY?

On Tue, May 30, 2006 at 09:06:48PM -0400, jarrod roberson
wrote:
> given that I have done .add_posting() correctly with the LP: prefixed terms
> and the second AND Query returns the set of data I expect, why does the
> first query NOT return ANYTHING?
I can't reproduce this.  The attached python script finds one match (and
also does using quartz or flint instead of the inmemory backend).

So it seems the circumstances required to trigger this are somewhat
subtle.  Can you produce a self-contained example which shows this
similar to phrase.py?
> also what will ulimate be "better" given the context of these
queries, the
> AND QUERY or the PHRASE QUERY?
I'm not sure I understand what you're trying to search for here.  But
generally a PHRASE query is better when the terms are only useful in
a particular order, or when there are common false positives which
includes all the terms but aren't relevant.

A phrase query is slower than an AND, so that's a reason to favour AND
if there's no other reason to pick one or the other.

Cheers,
    Olly
-------------- next part --------------
A non-text attachment was scrubbed...
Name: phrase.py
Type: text/x-python
Size: 599 bytes
Desc: not available
Url :
http://lists.tartarus.org/pipermail/xapian-discuss/attachments/20060601/37fd1833/phrase.py