jonnyfrastovner

You could also try to hack the HTML exporter source code if you have the skills in-house. This is rather easy, since font size is defined in only one place in the html exporter class. Simply replace SIZE_UNIT with the string "pt" where font size is defined, and html output will resemble ireport, PDF and word. In JR, you are forced to choose between points and pixels as size units for all elements (fonts, margins, table size etc), but AFAIK, the limitation only affects the HTML exporter, since you cannot express font sizes in pixels in excel, word, pdf, the internal ireport previewer etc, and you can't express box sizes in points in the same document formats (i believe) - you can only do that in HTML. This makes JR html output very different from output in other formats. I filed a bug, since it makes web output ugly and unreadable, but Jaspersoft has stated that the behaviour is by design.

I know. The issue is that crosstable cells are identified as frames, and therefore, the html export wraps the cell content (usually a textbox) in a table tag. However, the "frames as nested tables" feature is very useful in html reports because the page structure becomes much cleaner, particularly when using crosstabs. I still think that crosstab cells should never cause JR to create table tags. The solution is to define crosstab cells as something distinct from ordinary user-drawn frames (that is supposed to generate a table when using the nested tables option). An extra feature that many users would love, would be to give crosstab cells attributes identifying them as odd or even row/column. That would make a standard feature such as alternating row colors so much easier to implement for end users.

I think all the redundant code is indeed the problem. In addition, the extra table tag for each td makes crosstables look very bad in web pages. I filed a bug and a patch for this last year, but i dont believe web output is the highest priority in JR.

You can find an example <a href="http://dbh.nsd.uib.no/rapporter/nokkeltall/studenter_visning.action?inst=[institusjon.institusjoner%20flat%20liste].[alle%20institusjoner].[universitetet%20i%20Oslo]&Fagfelt=[Norsk%20standard%20for%20utdanningsgruppering%20(NUS)].[Totalt]&niva=1&nuskode=&nusniva=0&instkode=1110"> here </a> (sorry, norwegian only, drill using the column to the right). It was non-trivial to accomplish, i'd suggest you buy the ultimate guide, that helped us a lot. We used xmla rather than mondrian directly.

I posted the code as requested, and assigned it to you.

I would say this is a bug in the HTML exporter. Other formats use points as font metric, but the HTML exporter uses the report default size unit (usually defined in pixels). If you choose point as report size unit, table sizes etc become way to large, and again deviates from other output formats. Our solution for our web project deployment has been to modify and override the HTML exporter class (JRHtmlExporter.java), and force JR to use points as font metric and pixels otherwise. IMO this should be the default behavior for HTML output. Won't show correctly in ireport though, unless you modify and build jasperreports.jar from source. To much hassle for me, particularily with the frequent upgrades.

Using the option "export frames as nested tables", every cell in a crosstab is exported twice. Expected behaviour is that the crosstab should be embedded in a table cell as an individual table. The exporter outputs every cell in the crosstab as an individual table, and then exports the contents of the cell as a text field within the table. This produces a double set of boxes for every cell in the resulting web page, making output look bad, and documents potentially very big. Relevant calls as far as i can see is in JRHtmlExporter.java exportgrid method lines 804-828 (JR 3.1) Output from a 2x2 crosstab generated in iReport is attached. Code:JRPrintElement element = gridCell.getWrapper().getElement(); if (element instanceof JRPrintLine) { //several conditions omitted } else if (element instanceof JRPrintText) { exportText((JRPrintText)element, gridCell); } else if (element instanceof JRPrintFrame) { exportFrame((JRPrintFrame) element, gridCell); }

According to this press release, you should be able to connect to jasperanalysis cubes from excel. However, i get an error when trying to connect from excel 2007 ("element xmlschema not allowed under root") - this is the same error i get under standalone mondrian. Is the excel integration only available in the "professional" version? I dont find any mention of it in either the press release or any other places.

I believe you can not list two hierarchies of the same dimension at once in an mdx query. Anyways, in order to select a particular hierarchy other than the default, you have to name it explicitly in the schema and in the query. In the schema, add a name attribute to the second hierarchy, in the query, add a dot and the hierarchy name to the dimension, like this (assuming you name the second hierarcy "beschwerdeinhalt"): select [x] on columns, [Daten Feld.beschwerdeinhalt].[All beschwerdeinhalt] on rows In this case, you probably will have to make two separate dimensions beschwerdeinhalt and entschaedigungsform.

Hi, I am facing performance issues in an xpath query. The documents are structured as defined in the attached xsd. My testfile is about 2MB - 3000 nodes.I define the query as descendant::FeilMelding The report query window instantly recognizes 3000 nodes. I define two fields, as @type and @grad. Filling a query outputting the two fields takes about 18 seconds, and adding new fields slows the query down further. I have defined sorting of the two fields in the "sort options" window, but it does not seem to impact performance much. Is this performance as expected? I have a rather fast pc. I have tried to define an alternative query executor, as suggested elsewhere: net.sf.jasperreports.xpath.executer.factory=net.sf.jasperreports.engine.util.xml.JaxenXPathExecuterFactory but ireport gives this error: net.sf.jasperreports.engine.JRException:Ânet.sf.jasperreports.engine.JRRuntimeException:Ânet.sf.jasperreports.engine.JRException:ÂClassÂ"net.sf.jasperreports.engine.util.xml.JaxenXPathExecuterFactory"ÂshouldÂbeÂcompatibleÂwithÂ"net.sf.jasperreports.engine.query.JRQueryExecuterFactory"

I guess you have an xmla provider running? I do it by defining three variables in the JR report itself: XMLA_URI the ws adress, eg. "http://www.cubedata.com/xmla" XMLA_CATALOG name of the catalog, eg "myCubes" XMLA_DATASOURCE eg. "Provider=Mondrian;DataSource=myCubes;" The last two parameters are configured in datasources.xml on the xmla server (in MSAS and mondrian at least) In iReport you need to configure an xmla datasource to design and run the reports.

OK. But i still get an error stating " Error: ; nested exception is: org.xml.sax.SAXParseException: White spaces are required between publicId and systemId." as described in this thread: http://www.jasperforge.org/index.php?option=com_joomlaboard&Itemid=215&func=view&id=24241&catid=9

At least normal enough that i experienced the same. Creating the WEB-INF directory took care of the problem, but now i got another sax-exception stating that "Processing instructions are not allowed within soap messages". Does anyone know what this means?

I've tried to ask this on the jasperreports forum, but ... I got reports running on iReport using xmla. Works great. Now i want to deploy my reports on jboss. JR queries the web services, witch returns soap document containing the requested data. I have had several errors filling and compiling the report on Jboss (4.0.4). After dealing with i18n issues, it complains "No such tuple ([Measures].[calculatedmeasures1].[null]", witch i guess is because the measure "calculatedmeasures1" has no value for a given member. But still: why does it work in iReport? Are there iReport classes I can copy to my webapp to get it working?

I get the feeling i am the only one who cares, but here it goes: I got JR working in my web app by manually editing the report file, setting <queryString language="mdx"> instead of <queryString language="xmla-mdx"> as iReport would have it. No clue as to why this would work in iReport but not elsewhere. Now the ws returns soap as expected (yay!). However, in filling the report, JR balks on national characters (æøå), probably because the ws returns an ISO encoded document, not UTF8. Again, not an issue in iReport. Any idea is welcome! Exception below: [sTDOUT] 15:57:41,343 ERROR [JROlapDataSource] Error parsing field mapping line 1:6: expecting ']', found 'å' at net.sf.jasperreports.olap.mapping.MappingLexer.nextToken(MappingLexer.java:196) at antlr.TokenBuffer.fill(TokenBuffer.java:69) at antlr.TokenBuffer.LA(TokenBuffer.java:80) at antlr.LLkParser.LA(LLkParser.java:52)

Hi, I get the following exception when trying to run a xmla report from a jsp page: Caused by: net.sf.jasperreports.engine.JRRuntimeException: net.sf.jasperreports.engine.JRException: No query executer factory class registered for xmla-mdx queries. Create a propery named net.sf.jasperreports.query.executer.factory.xmla-mdx. at net.sf.jasperreports.engine.design.JRDesignDataset.queryLanguageChanged(JRDesignDataset.java:842) at net.sf.jasperreports.engine.design.JRDesignDataset.setQuery(JRDesignDataset.java:413) at net.sf.jasperreports.engine.design.JasperDesign.setQuery(JasperDesign.java:631) Anyone have a clue? I am using 131. The report works in Ireport. regards, Bjug Bøyum

I should probably mention that i run my apps on jboss (normally with a jndi data source).

I have successfully managed to create reports in iReport using XMLA and MDX against mondrian cubes. Now I want to deploy the report to a web page, optionally also create a report with subreports using different data sources/cubes. How do i specify data sources for xmla? I guess i must reference the WS URI (eg. http://localhost:8080/xmla) but how do i specify other parameters (catalog/cube, username/password etc)? How do i pass this to subreports? regards, - Bjug Boyum

Me too. Tried running Ireport with JRE1.5_11, without success. jdcryans wrote: jdcryans wrote: Now that I have updated, I do see my catalog and my cube but when I go in the query designer (with the XMLA connection and query language set to "XMLA-MDX") I have the following message : "defaultCloseOperation must be one of : DO_NOTHING_ON_CLOSE, HIDE_ON_CLOSE, or DISPOSE_ON_CLOSE". Funny enough! Anyone else has this message? This bug should be easy to fix. Thx Yes I do