We are doing a J2EE application using Myfaces,spring,hibernate. Currently we are using FOP to generate reports in our project, but it seems its not that fast to create a report in FOP from fast experience. i havent used any other reporting tool and now considering to use Jasper or Cognos (as we already have the license). Requirements are basically for 100 Online reports which are dynamically genrated and might be stored in file system.all reports are in pdf formats and some will be in multiple languages.

can anyone suggest which one will be best tool to do the thing. FOP i got problem with writing huge xsl’s and creating xml’s using java.

Thanks
Shail

We are doing a J2EE application using Myfaces,spring,hibernate. Currently we are using FOP to generate reports in our project, but it seems its not that fast to create a report in FOP from fast experience. i havent used any other reporting tool and now considering to use Jasper or Cognos (as we already have the license). Requirements are basically for 100 Online reports which are dynamically genrated and might be stored in file system.all reports are in pdf formats and some will be in multiple languages.

can anyone suggest which one will be best tool to do the thing. FOP i got problem with writing huge xsl’s and creating xml’s using java.

Thanks
Shail

thanks in advace,
Kashif

Would you recommend JFreeReport over JasperReports or BIRT to OpenReports and Pentaho users? If so, for what reasons?

Which engine would I recommend? Well, it depends.

If you need an embedded engine to provide printing capabilites in your application, I surely would head for JFreeReport. No other engine has a lower footprint and we do not touch the filesystem for the report processing. (And it is still no problem to create a 20.000 pages report within the default 64MB heap).

Those users are a bit more concerned about the performance than the common BI users. (A BI user simply buys a bigger maschine, but you cant do that for an accounting software, for instance.) In its smallest deployment, JFreeReport eats about 2MB, and noone has to recompile or fire up an IDE to strip not needed functionality from the engine. Embedded systems is where size matters.

For BI systems (on the Pentaho-plattform, for instance) or any other server deployment, I usually tell my users: Try JFreeReport, JasperReports and BiRT and use what fits your requirements most. Users rarely choose BIRT, and whether they choose JFreeReport or JasperReports depends on whether they used one of the tools before and (of course) their requirements.

(And most of the time, you could use a dice to predict the outcome. JasperReports and JFreeReport are very similiar in their functionality, we have some strength, JasperReports has others, but those differences are rather small.

Luckily, on either Pentaho or OpenReports, you dont have no vender lock-in, so you can use JFreeReport for one report and JasperReports for an other, if you wanted to. And if you decided for one tool and notice it does not cover that particular requirement, switch for that report. The freedom to choose does not vanish once you deployed your first report.

And finally:

Real world reports are also typically MUCH larger than the sample run.

I disagree on that one. I’ve met too many people how use the reporting engine to spit out thousands of one or two page reports (invoices, summaries, etc). For them, it is important that the engine performs well. They dont care about the capability to be faster on 10.000 page reports, they want to see results without having to wait. And for interactive use, 10 seconds more or less are noticable (and noticed!).

Regards,
Thomas

I will take a look at Crystal Reports XI. The engine may be free in some cases, but I don’t think the licensing terms allow distribution of the engine with OpenReports.

- Erik

I have not used iReport 1.2.7, beyond taking a quick look at the user interface. I was basing my opinion on iReport 1.2.5.

I agree that the performance is rarely the deciding factor when choosing a reporting solution. If it was, I don’t think it would rely on Galina and Valeria’s blog for my performance testing.

I would like to note that, although I am probably most familiar with and slightly biased towards JasperReports, it doesn’t really matter to me what reporting engine a person uses with OpenReports. OpenReports supports multiple reporting engines to provide users a choice, and I agree with you that prospective users should test multiple engines and decide which one best fits their requirements.

- Erik

Thomas, the blog entry you linked to with performance data doesn’t seem to reflect real world usage. I talk to our users all the time, and they are rarely doing a simple table dump of a small data set. They usually have a mix of reports, ranging from simple to highly complex, all running in the same server. Report authors spend most of their effort working on the look and feel of a report. They take advantage of the flexibility of a banded layout, use several fonts and colors, and often have a summary section with multiple charts, sometimes pulling data from multiple datasources. Real world reports are also typically MUCH larger than the sample run.

More importantly, I have found that performance is rarely the main factor people use to choose a reporting engine, or even a primary factor. The main influencer is the ability to create the reports that they want, with the look they want. The decision maker could be looking for crosstab report or barcode support, and will make their decision based on that. Or maybe its the ability to run a 10,000 page report without crashing. The important point is that different people have different needs, and will use different criteria.

I have a bias towards JasperReport (I’m a contributer and the CTO of JasperSoft) and Thomas has a bias towards JFreeReport - perfectly understandable. I suggest folks take a look at both and decide for themselves.

-Barry Klawans