Hello.
I checked out Quip 1.6.1 version via Java API . That version seemed good to use and was provided with Java API documentation.
As we tested new 2.1.1 version via SAG QuipGUI it seems more promisable, but ufortunately API specification is not delivered with it. As we explored Quip.jar it differs from previous version.
Where should i find API docs?
Regards
Novak Leonid
novak@ispras.ru
Sorry,
javadoc of the quipAPI got lost in the latest
distribution of Quip. However. the JavaAPI should
be usable in the same way as in the previous version of quip. Only small changes were made in
this part of quip. So most things should still
be valid from the documentation of quip 1.6.1.
Did you encounter any problem when you tried to
use the javaAPI the same way as with quip1.6.1?
Sven Eric
Hello.
That code doesn’t work with current version
(but worked with previous):
QuiPConnection qpc;
public QuipFileSystemConnection(String Path) throws XQueryException
{
qpc=new QuiPConnection(Path,null);
//Path is URI file system directory
}
Bat afterwords, as i understood , that there is no QuiPConnection class i changed the code slightly:
Connection qpc;
public QuipFileSystemConnection(String Path) throws XQueryException
{
Properties obj = new Properties();
obj.put(“home”,“C:/” );
obj.put(“quipcmd”,“./ru/server/quip.exe” );
qpc=DriverManager.getConnection(Path ,obj);
}
Now everything works fine, but still we have some problems:
1)Suppose i run my program from c:\test1\ directory, and file “test.xml” is placed in c:\test2\ directory. In previous version we initialized Path variable with “file://c:/test2/” string . And everything was good.
In this version it doesn’t work: (“QuipError:file not found”) (we also tried to use “home” variable). Now we need to rewrite query:
document(“c:\test2\test.xml”) instead of document(“test.xml”)
2)Here is the code of execution of query:
public synchronized String executeToString(String xquery) throws Exception
{
QueryResult qres =qpc.executeQuery(xquery);
String res=qres.getRawString();
return res;
}
We get an answer to our query with a huge amount of “space symbols”. And when we parse the answer we get additional text nodes, which contain only space symbols.
3) when we tried to query “really big” xml file (10 MB) we received an error:
“System error: quip.exe: fatal error: Windows programs can only use 256Mb of heap; sorry!”
What does it mean? As i know (MSDN library) Windows programs can use more than 256Mb of heap
Regards, Leonid Novak novak@ispras.ru
Hello.
That code doesn’t work with current version
(but worked with previous):
QuiPConnection qpc;
public QuipFileSystemConnection(String Path) throws XQueryException
{
qpc=new QuiPConnection(Path,null);
//Path is URI file system directory
}
Bat afterwords, as i understood , that there is no QuiPConnection class i changed the code slightly:
Connection qpc;
public QuipFileSystemConnection(String Path) throws XQueryException
{
Properties obj = new Properties();
obj.put(“home”,“C:/” );
obj.put(“quipcmd”,“./ru/server/quip.exe” );
qpc=DriverManager.getConnection(Path ,obj);
}
Now everything works fine, but still we have some problems:
1)Suppose i run my program from c:\test1\ directory, and file “test.xml” is placed in c:\test2\ directory. In previous version we initialized Path variable with “file://c:/test2/” string . And everything was good.
In this version it doesn’t work: (“QuipError:file not found”) (we also tried to use “home” variable). Now we need to rewrite query:
document(%
1)
It slipped our attention that the API has changed
in some respects and the javaDoc got lost in our
documentation folder. I am sorry for that.
There are no longer the QuipConnection classes but
these have been replaced by the
QMachineConnection. But you do not need to bother,
if you use the DriverManager.
The following class shows how to use this and
how to set the base drectory for your queries:
import java.util.;
import com.softwareag.xtools.quip.xqueryAPI.;
public class TestAPI {
public static void main(String args){
Properties prop = new Properties();
try {
prop.put(“quipcmd”,“quip.exe” );
Connection conex
= DriverManager.getConnection( “file:c:\cygwin\home\”, prop );
QueryResult res = conex.executeQuery((String)args[0]);
System.out.println(res.getRawString());
}
catch (XQueryException e) {
System.out.println(“XQueryException: " + e.getMessage());
}
}
}
noltic that the uri in the connection is of the
form:
file:c:/…
so in this case no slashes after the first colon.
2) You noticed the whitespaces. The current quip
version keeps all whitespaces. They are
text node constructors. Actually whitespace
is a rather nasty thing. So we better keep them.
These are text-nodes.
You can avoid unwanted white-spaces
in the way you layout your query:
write:
{
and write
<a
><b
/>
3) on windows the quip.exe cannot exceed 256M
during execution. The error message you get
is actually produced by the Haskell runtime
system of ghc (Home — The Glasgow Haskell Compiler)
You can avoid this message if you set the maximum
heap size of the quip.exe to some smaller value
by the command line argument:
+RTS -M200M -RTS
In the API you set this value as command line
property:
prop.put(“qmachineoptions”,”+RTS -M200M -RTS");
However this might end up in the message:
System error: Heap exhausted
hope this helps
Sven Eric
Hello.
Thank you very much for answers to my questions. The answer to question 1) (concerning path property) really helped us, but other questions are still open:
We have two xml files: test1.xml and test2.xml. And a query:
FOR $x in document(“test1.xml”)/* RETURN $x
test1.xml is 1MB file , and test2.xml is 10MB file
Also they contain no whitespaces
(e.g hh…)
1) After we execute this query under Quip we receive an answer with whitespaces, while initial document didn’t contain them. Moreover we have two line breaks after each element node.
2)when we run FOR $x in document(“test2.xml”)/* RETURN $x
we receive “not enough memory” error. As you explained this is Haskell error (Heap size is set to 256M). Can you tell me what is the maximum size of xml document, that can be queried with quip engine.
Regards, Leonid Novak novak@ispras.ru
You are right, Quip layouts its result and adds
whitespaces. However only whitespaces which are
not significant, which are so to speak allowed
to add. (unless you prove the opposite, which
would mean that there is a bug)
It is hard to tell, which the maximum doc size
is you can process with quip. (on linux this is
only restricted on your hardware memory). It
depends what you are doing with the document.
BTW: you should set the heap size a bit smaller
than 256 Mb, beacuse the quip.exe allocates some
more memory (for stack etc.) and the complete
amouint may not exceed 256mb. (it might be that
this restriction is dropped from the Haskell guys
some day…)
Sven Eric
Can the Documentation be made available by some means ??
Regards,
Amit