Hi,
I am trying to open a collection of files where some of them contain character entity references like – which are not declared in the document and so make basex crash if I use collection()
If I use file:read-text for each file and then put this functioni in file:write, the saved text will show < > instead of < and >: why?
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db usind db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml. I have tried to declare other db options in the same way but they seem not be read.
Giuseppe
Hi,
I forward a message I sent a few days ago which apparently did not receive any answer. I am trying to open a collection of files where some of them contain character entity references like – which are not declared in the document and so make basex crash if I use collection()
If I use file:read-text for each file I can open them, but if I then put this functioni in file:write, the saved text will show < > instead of < and >: why?
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db using db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml (which means that the declaration at the beginning of the query is not read). I have tried to declare other db options in the same way (ie in the prolog of a script) but they seem not be read.
Giuseppe
Hello Joseph,
sorry, I didn't have much time to reply to your mail before. So, using file:write without any optimization parameters it will most likely try to serialize as XML and then it should replace e.g. < by <. Instead, you could use file:write-text() or set the seriaziation method explicitely, i.e. using file:write(<path>, <data>, map { "method": "text" })
Using skipcorrupt should indeed work. Could you please send us the concrete snippet you are trying to execute? And also, which BaseX version are you using?
Cheers Dirk
On 09/03/2015 01:05 PM, meumapple wrote:
Hi,
I forward a message I sent a few days ago which apparently did not receive any answer. I am trying to open a collection of files where some of them contain character entity references like – which are not declared in the document and so make basex crash if I use collection()
If I use file:read-text for each file I can open them, but if I then put this functioni in file:write, the saved text will show < > instead of < and >: why?
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db using db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml (which means that the declaration at the beginning of the query is not read). I have tried to declare other db options in the same way (ie in the prolog of a script) but they seem not be read.
Giuseppe
Hi Giuseppe,
I am trying to open a collection of files where some of them contain character entity references like – which are not declared in the document and so make basex crash if I use collection()
Could you please give us more information what "basex crash" means in this context?
If I use file:read-text for each file and then put this functioni in file:write, the saved text will show < > instead of < and >: why?
Dirk already gave you the answer I think. Here are two possible alternatives:
try { doc(...) } catch * { () (: file is corrupt :) }
try { parse-xml(file:read-text(...)) } catch * { () (: file is corrupt :) }
Best, Christian
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db usind db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml. I have tried to declare other db options in the same way but they seem not be read.
Giuseppe
I forgot to give feedback to your last comment:
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db usind db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml. I have tried to declare other db options in the same way but they seem not be read.
Thanks for pointing this out. The examples in the Wiki were outdated; I have just updated them. The way to do it now is to specify the options as function arguments.
Christian
So...you mean that "declare db:skipcorrupt ... " is not a valid syntax anymore?
Il giorno 14/set/2015, alle ore 15:44, Christian Grün christian.gruen@gmail.com ha scritto:
I forgot to give feedback to your last comment:
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db usind db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml. I have tried to declare other db options in the same way but they seem not be read.
Thanks for pointing this out. The examples in the Wiki were outdated; I have just updated them. The way to do it now is to specify the options as function arguments.
Christian
So...you mean that "declare db:skipcorrupt ... " is not a valid syntax anymore?
The syntax "declare option ..." is still valid, but many options don't apply to functions anymore. As it's not clearly documented what options are interpreted by XQuery functions, it may be that we will completely remove global BaseX options in future, or at least restrict them to a well-defined subset.
Il giorno 14/set/2015, alle ore 15:44, Christian Grün christian.gruen@gmail.com ha scritto:
I forgot to give feedback to your last comment:
Moreover, if I create a database from the GUI and select the option "skip corrupt", the database is created. However if I follow instruction here http://docs.basex.org/wiki/Parsers In order to add single files to a db usind db:add and precede the query with declare option db:skipcorrupt 'true', the files are not added because I get an error message due to the not well-formed xml. I have tried to declare other db options in the same way but they seem not be read.
Thanks for pointing this out. The examples in the Wiki were outdated; I have just updated them. The way to do it now is to specify the options as function arguments.
Christian
basex-talk@mailman.uni-konstanz.de
-
Christian Grün -
Dirk Kirsten -
meumapple