database corruption after update query

Monday, 19 December 2011 · 4 messages · last reply 20 Dec 2011
AH
Adrian Hossu19 Dec 2011 10:26 #
Hello all, first the info: basex 7.0.2 debian 64bit jre 1.6.0_30-b12 We are experiencing some strange behaviour with basex - certain update queries are causing our database to become corrupted, and we need to drop the database and re-import in order to be able to continue. Here is an example of such a query: replace node collection("ft")/user/user[@id='2']/logsheets/logsheet[@date='2011-12-20']/exercise[2] with <exercise/> The query executes successfully, however, executing db:list('ft') after this call gives us only garbled text. The basexgui shows the results of this as: &#x0D;&#x16;&gt;&#x95;&#x04;_'蘼&#x01; &#x0D;&#x16 ..................etc.... Perhaps it's the query, or our original data which we are starting with? Any help would be greatly appreciated. best regards, Adrian
CG
Christian Grün19 Dec 2011 17:51 #
Dear Adrian, thanks for reporting this issue, which should be fixed as soon as possible. First of all: does it also occur with the latest snapshot (7.1 beta, [1])? If yes, do you think you can create an sccee [2] that allows us to reproduce the problem? Christian [1] http://files.basex.org/releases/latest [2] http://sscce.org
Quoted text · Adrian Hossu · 26 lines
On Mon, Dec 19, 2011 at 10:26 AM, Adrian Hossu <adibadi at gmail.com> wrote: > Hello all, > > first the info: > basex 7.0.2 > debian 64bit > jre 1.6.0_30-b12 > > We are experiencing some strange behaviour with basex - certain update > queries are causing our database to become corrupted, and we need to drop > the database and re-import in order to be able to continue. Here is an > example of such a query: > > replace node > collection("ft")/user/user[@id='2']/logsheets/logsheet[@date='2011-12-20']/exercise[2] > with <exercise/> > > The query executes successfully, however, executing db:list('ft') after this > call gives us only garbled text. The basexgui shows the results of this as: > &#x0D;&#x16;&gt;&#x95;&#x04;_'蘼&#x01; &#x0D;&#x16 ..................etc.... > Perhaps it's the query, or our original data which we are starting with? > > Any help would be greatly appreciated. > > best regards, > Adrian
CG
Christian Grün20 Dec 2011 11:42 #
Hi Adrian, I'm glad to report that your updating problem is fixed. The bug occurred in the incremental document index (i.e., the database itself was still consistent). Please have a look at the latest, stable snapshot (the next official release is due to Jan 12 next year): http://files.basex.org/releases/latest/ Hope this helps, Christian
Quoted text · Christian Grün · 41 lines
On Mon, Dec 19, 2011 at 5:51 PM, Christian Grün <christian.gruen at gmail.com> wrote: > Dear Adrian, > > thanks for reporting this issue, which should be fixed as soon as > possible. First of all: does it also occur with the latest snapshot > (7.1 beta, [1])? If yes, do you think you can create an sccee [2] that > allows us to reproduce the problem? > > Christian > > [1] http://files.basex.org/releases/latest > [2] http://sscce.org > ___________________________ > > On Mon, Dec 19, 2011 at 10:26 AM, Adrian Hossu <adibadi at gmail.com> wrote: >> Hello all, >> >> first the info: >> basex 7.0.2 >> debian 64bit >> jre 1.6.0_30-b12 >> >> We are experiencing some strange behaviour with basex - certain update >> queries are causing our database to become corrupted, and we need to drop >> the database and re-import in order to be able to continue. Here is an >> example of such a query: >> >> replace node >> collection("ft")/user/user[@id='2']/logsheets/logsheet[@date='2011-12-20']/exercise[2] >> with <exercise/> >> >> The query executes successfully, however, executing db:list('ft') after this >> call gives us only garbled text. The basexgui shows the results of this as: >> &#x0D;&#x16;&gt;&#x95;&#x04;_'蘼&#x01; &#x0D;&#x16 ..................etc.... >> Perhaps it's the query, or our original data which we are starting with? >> >> Any help would be greatly appreciated. >> >> best regards, >> Adrian
AH
Adrian Hossuin reply to Christian Grün, 19 Dec 2011, 17:5120 Dec 2011 14:15 #
Thanks to Christian, this issue has been resolved in the build from 2011.12.20. http://files.basex.org/releases/latest/ - Adrian On Mon, Dec 19, 2011 at 5:51 PM, Christian Grün <christian.gruen at gmail.com>wrote:
Quoted text · Christian Grün · 45 lines
> Dear Adrian, > > thanks for reporting this issue, which should be fixed as soon as > possible. First of all: does it also occur with the latest snapshot > (7.1 beta, [1])? If yes, do you think you can create an sccee [2] that > allows us to reproduce the problem? > > Christian > > [1] http://files.basex.org/releases/latest > [2] http://sscce.org > ___________________________ > > On Mon, Dec 19, 2011 at 10:26 AM, Adrian Hossu <adibadi at gmail.com> wrote: > > Hello all, > > > > first the info: > > basex 7.0.2 > > debian 64bit > > jre 1.6.0_30-b12 > > > > We are experiencing some strange behaviour with basex - certain update > > queries are causing our database to become corrupted, and we need to drop > > the database and re-import in order to be able to continue. Here is an > > example of such a query: > > > > replace node > > > collection("ft")/user/user[@id='2']/logsheets/logsheet[@date='2011-12-20']/exercise[2] > > with <exercise/> > > > > The query executes successfully, however, executing db:list('ft') after > this > > call gives us only garbled text. The basexgui shows the results of this > as: > > &#x0D;&#x16;&gt;&#x95;&#x04;_'蘼&#x01; &#x0D;&#x16 > ..................etc.... > > Perhaps it's the query, or our original data which we are starting > with? > > > > Any help would be greatly appreciated. > > > > best regards, > > Adrian >

⚡Generated with XQuery