Page 128 - ASBIRES-2017_Preceedings

Thenuwara & Munasinghe

4 DATA COLLECTION AND using a different tool (SQL developer) to
ANLYSIS fetch the data from the database view,
totally refraining the use of Crystal filtering.
As mentioned in the Methodology The observation was not considerably
section, a comparison between before and different.
after response times was done. The reports
with the lowest performance have been As suggested in literature review
considered here, as the other reports were section regarding the use of a view instead
responding less than one minute after the of a table could be the cause behind this
changes has been made. Figure 1 shows the tool-independent behaviour, a table (more
variation of the time to generate the selected likely an instance of a view) was created
reports before and after the enhancements. from the view using a query to use with this
particular report and tested for performance
25 increase.
20
T Performan As a result, a quick increment in
i 15 ce Before performance was observed. However, this
m 10 solution could not be adhered straightway
Performan because converting views into tables could
e 5
ce After not be done at the overall database level. It
0 was observed that, application of such a
1 2 3 4 5 6 7 8 solution could cost the developer’s overhead

Figure 1: Performance of the reports of time and may raise security issues at the
selected database level as, unlike views which
restrict users of direct accessing the data,
Figure 2 explains how the number of providing a table could open the database
pages varied with the time for the selected for the client to access
reports.
6 CONCLUSION
300
Performance is an important factor to
250 Perfo
rman maintain quality of a product and to deliver
200 ce good value added services. This is important
Befor
150 e in report generation as it is involved in
Pages
analysis and prediction of statistical data. To
100
Perfo improve performance of a report generation
50 rman process, process time should be decreased to
ce
0 After a possible minimum. There are many
1 2 3 4 5 6 7 8 alternative solutions to increase report

generation speed in Chrystal Reports and
Figure 2: Performance of the reports among them the re-formatting and
selected enhancing the filtering process of Crystal
syntaxes can be identified as the most
5 RESULT AND DISCUSSION convenient and cost effective solution. In
this research, the application of this showed
As shown in the evaluation in section a considerable progress regarding the
4, the performance of the report generation identified slow report generation issue.
has been increased remarkably after Moreover, there can be more advanced
applying the solutions. Therefore, it proves solutions, which can be used if the problem
that filtering has a considerable impact on cannot be resolved only by enhancing the
the performance.
filtering. Since crystal sends all criteria to
All the slow performing reports could the database in one query, by pre-processing
be resolved by the proposed solution with and filtering complex data the fundamental
the exception of one. For that particular weakness of Crystal's one-query approach
report a further investigation was conducted can be overcome in report creation.

118