Bugs item #2810280, was opened at 2009-06-22 15:45
Message generated for change (Settings changed) made by hrode
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2810280&group_id=56967
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: PFtijah
Group: Pathfinder "stable"
>Status: Closed
>Resolution: Works For Me
Priority: 5
Private: No
Submitted By: francois guerin (fguerin)
Assigned to: Henning Rode (hrode)
Summary: PFtijah: full text index exists or not exists.
Initial Comment:
Hi,
the mapi client keeps on telling me two contradictory things:
- when i try to create an index (it seems that indexes are automaticaly created
with module(pathfinder), Mapi tells me that an index has already been created.
- when i try to make a full text query, Mapi tells that there's no index
created.
$ ./bin/mclient -lx --time
xquery>tijah:create-ft-index()
more><>
MAPI = mone...@localhost:50000
QUERY = tijah:create-ft-index()
ERROR = !ERROR: tj_init_collection, pftijah collection already exists:
DFLT_FT_INDEX
Timer 11.846 msec
xquery>let $c := collection("frwikipedia")
for $res in tijah:query($c, "//T[about(.,marcher)]")
return $res/Tranche/Sentence/T
more><>
MAPI = mone...@localhost:50000
QUERY =
ERROR = !ERROR: tj_get_ft_index, ft-index "DFLT_FT_INDEX" does not exist.
create full text index using tijah:create-ft-index()
Timer 22.598 msec
xquery>
(MacBook - Monetdb MAY release - 32bits - installation: --enable-xquery
--enable-optimise)
----------------------------------------------------------------------
>Comment By: Henning Rode (hrode)
Date: 2009-09-15 17:21
Message:
test case cannot be added. the main problem here is already that the error
is not reproducable. the query mentioned in the bug reports works fine for
me, and is already represented by other test cases.
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2009-08-21 11:49
Message:
Is there a test for this one?
If so, please state where;
otherwise, please (try to) add one, or explain why that isn't (easily)
possible.
After that, please close, again.
Thanks!
----------------------------------------------------------------------
Comment By: Henning Rode (hrode)
Date: 2009-08-20 16:24
Message:
the above described indexing procedure works completely fine here, even on
your large dataset.
the query indeed fails on my machine as well, but with a completely
different and reasonable error message asking you to use either a relative
path expression in your NEXI query or to use a tijah:query without a set of
context nodes.
i suppose your error comes is the result of previous experiments resulting
in a corrupt database. so, maybe you should remove the entire database and
then restart the experiment. it is also possible that indexing this large
dataset does simply not work on your laptop, since you probably need here
main memory >4GB
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2009-08-14 14:01
Message:
tagged subject
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2009-08-11 15:19
Message:
Thanks, Francois!
... the ball is back with Henning ;-) ...
----------------------------------------------------------------------
Comment By: francois guerin (fguerin)
Date: 2009-08-11 15:15
Message:
how to reproduce the problem:
- dl the archive http://fguerin.inria.perso.neuf.fr/final_wiki.tar.bz2
- unzip it
- store xml files under a collection called "frwikipedia" (read-only)
- then, do what I have described in the first post.
and voila, here comes the error.
----------------------------------------------------------------------
Comment By: Stefan Manegold (stmane)
Date: 2009-08-11 13:34
Message:
Francois,
could you please provide us with detailed instructions how to re-produce
your problem?
Thanks!
Stefan
----------------------------------------------------------------------
Comment By: Henning Rode (hrode)
Date: 2009-08-05 15:31
Message:
I never got any answer from the bug submitter. I left it open at that time,
because I had seen similar problems before and was not sure whether they
are solved in the completely current version. However, I cannot do anything
here without being able to re-produce the problem.
----------------------------------------------------------------------
Comment By: Sjoerd Mullender (sjoerd)
Date: 2009-08-05 15:21
Message:
What is the status?
----------------------------------------------------------------------
Comment By: Henning Rode (hrode)
Date: 2009-06-30 09:33
Message:
i changed a few checks in the code that are applied at indexing time.
if that does not help here, i would really need a script to reproduce the
bug.
so, without further input, i will close the bug in a few weeks.
----------------------------------------------------------------------
Comment By: Henning Rode (hrode)
Date: 2009-06-23 15:40
Message:
hej francois,
i have seen similar problems before in old versions of pf/tijah, when the
index got inconsistent due to commit statements at wrong places in the
internal indexing procedure. most people simply emptied the database
completely and re-indexed everything, which usually works fine then, but it
is obviously not a solution.
now, the only problem is, that i need some way to reconstruct your error
here. can you describe in a few steps (xqueries) what you did. i would need
some xquery that is reproduceable here and leads to the same error
message.
best -henning
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2810280&group_id=56967
------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
